10.4324 9781315800820 Previewpdf

Self-Efficacy
This page intentionally left blank

SELF -EFFICACY:
THOUGHT CONTROL
OF ACTION
Edited by
Ralf Schwarzer
Freie Universitat Berlin
~~o~:~~n~s~~up
Routledge
Routledg
LONDON AND NEW YORK

First published 1992 by
Taylor & Francis Group
This edition published 2014 by Routledge

2 Park Square, Milton Park, Abingdon, Oxon OXl4 4RN
711 Third Avenue, New York, NY 10017, USA
Routledge is an imprint o/the Taylor & Francis Group, an informa business
SELF-EFFICACY: Thought Control of Action
Copyright © 1992 by Hemisphere Publishing Corporation. All rights reserved.

Except as permitted under the United States Copyright Act of 1976, no part of this
publication may be reproduced or distributed in any form or by any means, or stored
in a database or retrieal system, without the prior written permission of the publisher.
Cover design by Michelle Fleitz.

A CIP catalog record for this book is available from the British Library.
Library of Congress Cataloging-in-Publication Data

Self-efficacy : thought control of action / edited by Ralf Schwarzer.
p. cm.
lncludes bibliographical references and indexes.
I. Control (Psychology) 2. Self-perception. 3. Action theory.
4. Medicine and psychology. I. Schwarzer, Ralf.
[DNLM: 1. Adaptation, Psychological. 2. lnternal-External
Control. 3. Self Concept. BF 697 S4653)
BF61l.S43 1992
153.8-dc20
DNLMIDLC
for Library of Congress 92-1465
ISBN 1-56032-269-1 CIP
CONTENTS
Contributors vii
Preface ix
I Self-Efficacy and Human Functioning
Exercise of Personal Agency Through the Self-Efficacy Mechanism

Albert Bandura 3
Two Dimensions of Perceived Self-Efficacy: Cognitive Control
and Behavioral Coping Ability
William McCarthy & Michael Newcomb 39
Expectancies as Mediators Between Recipient Characteristics
and Social Support Intentions
Ralf Schwarzer, Christine Dunkel-Schetter, Bernard Weiner,
& Grace Woo 65
II Self-Efficacy and Human Development
Perceived Control: Motivation, Coping, and Development

Ellen Skinner 91
Adults Expectancies About Development and its Controllability:
Enhancing Self-Efficacy by Social Comparison
Jutta Heckhausen 107
Personal Control Over Development: Implications of
Self-Efficacy
Jochen Brandtstadter 127
III Self-Efficacy, Stress, and Emotions
Perceived Self-Efficacy and Phobic Disability

S. Lloyd Williams 149
Self-Efficacy and Depression
David Kavanagh 177
Self-Efficacy as a Resource Factor in Stress Appraisal
Processes
Matthias Jerusalem & Ralf Schwarzer 195
vi Contents
IV Self-Efficacy and Health Behaviors
Self-Efficacy in the Adoption and Maintenance of Health

Behaviors: Theoretical Approaches and a New Model
Ralf Schwarzer 217
Self-Efficacy and Attribution Theory in Health Education
Gerjo Kok, Dirk-Jan Den Boer, Hein De Vries, Frans Gerards,
Harm J. Hospers, & Aart N. Mudde 245
The Influence of Expectancies and Problem-Solving Strategies
on Smoking Intentions
Martin V. Covington & Carol L. Omelich 263
V Self-Efficacy, Physical Symptoms, and Rehabilitation of

Chronic Disease
The Role of Physical Self-Efficacy in Recovery From

Heart Attack
Craig K. Ewart 287
Perceived Self-Efficacy in Self-Management of Chronic
Disease
Halsted R. Holman & Kate Lorig 305
Self-Efficacy Expectancies in Chronic Obstructive
Pulmonary Disease Rehabilitation
Michelle T. Toshima, Robert M. Kaplan, & Andrew L. Ries 325
Self-Efficacy Mechanism in Psychobiologic Functioning
Albert Bandura 355
Author Index 395

Subject Index 405
CONTRIBUTORS
Albert Bandura, Department of Psychology, Stanford University, Building 420,
Jordan Hall, Stanford, CA 94305, USA
Jochen Brandtstiidter, Fachbereich I-Psychologie, Universitiit Trier, Postfach 3825,
W-5500 Trier, Gennany
Martin V. Covington, Psychology Department, University of California at Berkeley,
3210 Tolman Hall, Berkeley, CA 94720, USA
Dirk-Jan Den Boer, Rijksuniversiteit Limburg, Department of Health Education,
P.O. Box 616, 6200 MD Maastricht, The Netherlands
Hein De Vries, Rijksuniversiteit Limburg, Department of Health Education,
Christine Dunkel-Schetter, Department of Psychology, University of California at
Los Angeles, Los Angeles, CA 90024-1563, USA
Craig K. Ewart, The Johns Hopkins University, Department of Behavioral Sciences &
Health Education, School of Hygiene and Public Health, 624 N. Broadway, Baltimore,
MD 21205, USA
Frans Gerards, Rijksuniversiteit Limburg, Department of Health Education,
Jutta Heckhausen, Max-Planck-Institut fUr Bildungsforschung, Lentzeallee 94,
W-lOOO Berlin 33, Germany
Halsted R. Holman, Stanford University Medical Center, Division of Immunology and
Rheumatology, 1000 Welch Road, Suite 203, Palo Alto. CA 94304, USA
Harm J. Hospers, Rijksuniversiteit Limburg, Department of Health Education,
Matthias Jerusalem, Institut fUr Psychologie (WE 7), Freie Universitat Berlin,
Habelschwerdter Allee 45, W- 1000 Berlin 33. Gennany
Robert M. Kaplan, Department of Community and Family Medicine, School of
Medicine, University of California at San Diego, 9500 Gilman Drive. La Jolla, CA
92093-0622, USA
David Kavanagh, Department of Psychology, University of Sydney, Sydney,
NSW 2006, Australia
Gerjo Kok, Rijksuniversiteit Limburg, Department of Health Education. P.O. Box 616,
6200 MD Maastricht, The Netherlands
Kate Lorig, Department of Medicine, Stanford University, 1000 Welch Road, Suite 320,
Palo Alto, CA 94304-1808, USA
William J. McCarthy, Department of Cancer Control, 1100 Glendon Avenue, Suite 711,
Los Angeles, CA 90024-1563, USA
viii Contributors
Aart N. Mudde, Rijksuniversiteit Limburg, Department of Health Education,

Michael D. Newcomb, Division of Counseling and Educational Psychology, WPH 500,
University of Southern California, Los Angeles, CA 90089-0031, USA
Carol L. Omelich, Psychology Department, University of California at Berkeley,
3210 Tolman Hall, Berkeley, CA 94720, USA
Andrew L. Ries, Department of Community and Family Medicine, School of Medicine,
University of California at San Diego, La Jolla, CA 92093-0622, USA
Ralf Schwarzer, Institut fUr Psychologie (WE 7), Freie Universitat Berlin,
Habelschwerdter Allee 45, W-I000 Berlin 33, Germany
Ellen Skinner, Graduate School of Education and Human Development, University of
Rochester, Rochester, NY 14627, USA
Michelle Toshima, Department of Rehabilitation Medicine, School of Medicine,
University of Washington, Seattle, WA 98109, USA
Bernard Weiner, Department of Psychology, University of California at Los Angeles,
S. Lloyd Williams, Department of Psychology, Lehigh University, Chandler-Ullmann
Hall #17, Bethlehem, PA 18015, USA
Grace Woo, Department of Psychology, University of California at Los Angeles,
PREFACE
Human functioning is facilitated by a personal sense of control. If people
believe that they can take action to solve a problem instrumentally, they become
more inclined to do so and feel more committed to this decision.
While outcome expectancies refer to the perception of the possible conse-
quences of one's action, self-efficacy expectancies refer to personal action
control or agency. A person who believes in being able to cause an event can
conduct a more active and self-determined life course. This "can do"-cognition
mirrors a sense of control over one's environment. It reflects the belief of being
able to control challenging environmental demands by means of taking adaptive
action. It can be regarded as a self-confident view of one's capability to deal with
certain life stressors.
Self-efficacy makes a difference in how people feel, think and act. In terms
of feeling, a low sense of self-efficacy is associated with depression, anxiety, and
helplessness. Such individuals also have low self-esteem and harbor pessimistic
thoughts about their accomplishments and personal development. In terms of
thinking, a strong sense of competence facilitates cognitive processes and aca-
demic performance. When it comes to preparing action, self-related cognitions
are a major ingredient of the motivation process. Self-efficacy levels can enhance
or impede motivation. People with high self-efficacy choose to perform more
challenging tasks. They set themselves higher goals and stick to them. Actions
are preshaped in thought, and people anticipate either optimistic or pessimistic
scenarios in line with their level of self-efficacy. Once an action has been taken,
high self-efficacious persons invest more effort and persist longer than those low
in self-efficacy. When setbacks occur, they recover more quickly and maintain
the commitment to their goals. Self-efficacy also allows people to select
challenging settings, explore their environments, or create new environments.
Self-efficacy is considered to be specific, that is, one can have more or less
firm beliefs in different domains of functioning. A sense of competence can be
acquired by mastery experience, vicarious experience, verbal persuasion, or phy-
siological feedback. Self-efficacy, however, is not the same as positive illusions
or unrealistic optimism. Since it is based on experience and does not lead to
unreasonable risk taking. Instead, it leads to venturesome behavior that is within
reach of one's capabilities.
Self-referent thought has become an issue that pervades psychological
research in many domains. It has been found that a strong sense of personal effi-
cacy is related to better health, higher achievement, and better social integration.
This concept has been applied to such diverse areas as school achievement, emo-
tional disorders, mental and physical health, career choice, and sociopolitical
change. It has become a key variable in clinical, educational, social, develop-
mental, health, and personality psychology.
x Preface
The present volume, organized in five sections, covers a broad range of

studies that deal with self-efficacy. The first section refers to self-efficacy and
human functioning.
In his introductory chapter, Albert Bandura outlines the role of self-efficacy
as part of his Social Cognitive Theory. Self-efficacy does not simply reflect the
perception of accomplishments; instead, it is based on subjective inferences from
different sources of information. Perceived self-efficacy has been shown to exert
an effect on performance, independent of actual ability levels. For example, self-
beliefs that were induced experimentally led to subsequent behavioral change. A
sense of efficacy is considered to influence different processes of human func-
tioning: cognitive processes, motivation, affect, and selection of environments.
William McCarthy and Michael Newcomb focus their chapter on the distinc-
tion of perceived self-efficacy either in terms of thought control or in terms of
actual performance. This parallels the common distinction between cognitive
coping ability and behavioral coping ability. In some situations, people may feel
competent to regulate themselves by reappraising the demands or by controlling
their emotions, whereas in other situations they may feel competent to change the
stressful encounters instrumentally. The authors studied the perceptions of a large
sample of young adults and found that issues such as purpose in life or loss of
control were only related to the cognitive control dimension, whereas social
stress issues such as assertiveness, leadership, and dating were only related to the
behavioral control dimension. These results may stimulate further research on
one's beliefs about self-regulation in cases where no explicit action is desired.
Social support has been found to buffer stress under certain circumstances.
Ralf Schwarzer, Christine Dunkel-Schetter, Bernard Weiner, and Grace Woo
have experimentally investigated conditions under which social support is ex-
tended. Victims who suffer from uncontrollable life events and invest high cop-
ing efforts are more likely to receive support, in contrast to those who experience
controllable events and do not cope well. Outcome expectancies and self-efficacy
expectancies emerged as mediators in the cognition-emotion-behavior pro-
cess. If the victim's condition was judged as changeable, and if the provider felt
competent to extend support, then the intent to help was most likely. This
research points to the possibility that people often do not behave in a supportive
way because they believe they are not capable of taking appropriate action.
The following three chapters add a developmental perspective in the second
section. Ellen Skinner reviews research on competence motivation and perceived
control in children and establishes a connection to the self-efficacy construct.
Self-efficacy expectancies are labeled "capacity beliefs" in her theoretical model,
as opposed to "strategy beliefs" or outcome expectancies, which are perceptions
of the behavioral causes of success or failure. Human beings are supposed to
strive for mastery and control over their environments and to attempt to generate
desired events. A need for competence, that may be innate and pervasive, moti-
vates behavior across the life span. Loss of perceived control can lead to coping
deficits and helplessness, particularly if this is due to self-efficacy impairment.
Preface xi
However, this is not so much the case if it is due to the interference of powerful
others.
Iutta Heckhausen explores the psychological mechanisms that alter and
maintain self-efficacy, particularly the role of infonnation from social compari-
son processes. Self-efficacy expectancies have to be realistic in order to translate
into successful actions and to avoid high risk-taking. On the other hand, positive
illusions set the stage for venturesome behavior and for an optimistic approach
toward life's challenges. The self-serving bias in beliefs about one's capabilities
and coping skills conveys motivational benefits but can be counterproductive
when the appropriate action either fails or is not available. Strategic social com-
parison can balance these conflicting demands. Age groups provide a frame of
reference for such comparisons. Goals that are congruent with those of one's ref-
erence group can be seen as realistic, whereas those that transcend this nonn can
be considered as self-enhancing.
Iochen Brandtstadter states that individuals contribute actively to shaping
their personal development and circumstances of living. This chapter focusses on
developmental implications of self-efficacy by using a large cohort study to
examine perceived control and coping preferences across the life-span.
Specifically, Brandtstadter studies how people deal with their perceived gains
and losses over the life span and how these are actively constructed. Personal
self-regulation changes with increasing age, reflected by a decline in tenacious
goal pursuit and an increase in flexible adjustment of developmental goals. This
is in line with a shift from a more instrumental or active-assimilative coping
mode to a more accomodative mode of coping.
Self-efficacy expectancy is inversely related to anxiety and depression and
can represent a powerful stress resource factor. This resource factor also allows
one to deal better with uncertainty, distress, and conflict. The third section of this
volume gathers articles that deal with stress and emotions and the way self-
efficacy interacts with them. S. Lloyd Williams reviews the literature on self-
beliefs as causal factors in phobia and finds that self-efficacy judgments are the
best predictors of therapeutic change. No other available theory provides such
strong evidence. He presents empirical research on severely phobic patients who
underwent psychological treatment. Those who were trained to master threaten-
ing situations and, by this, to build up a sense of competence, were more success-
ful than those exposed to other psychotherapeutic procedures.
David Kavanagh extends the scope of self-efficacy theory to the development
of depression and explores the predictive and causal influence that self-beliefs
have on the occurrence of depressive episodes. There appears to be a reciprocal
relationship between emotions and self-related cognitions. Depressive mood is
triggered by cognitions; emotions represent one source of infonnation, among
others, to shape self-efficacy. On the other hand, low self-efficacy deepens sad-
ness and increases vulnerability toward depressive episodes. This issue has been
examined by manipulating emotions experimentally. People in an induced sad
mood were reporting lower self-efficacy than those in a happy mood. In addition,
xii Preface
performance is related to both emotions and self-related cognitions. All three

constructs exert a direct effect on each other but also serve as mediators between
one another.
Matthias Jerusalem and Ralf Schwarzer attempt to integrate transactional
stress theory with social cognitive theory. They regard generalized self-efficacy
as one of the personal resource factors that counterbalance taxing environmental
demands in the stress appraisal process. Stress can be cognitively appraised as
either a challenge, threat, or harm/loss. In a laboratory experiment, subjects who
were confronted with difficult tasks under time pressure received fictitious per-
formance feedback. Those with high trait self-efficacy made more favorable
interpretations of the stressors compared to their low self-efficacy counterparts.
The authors conclude that dispositional self-efficacy not only facilitates coping
with stress but is already operating at an earlier phase of the stress process,
namely at the cognitive appraisal stage.
The fourth section relates self-beliefs to health behaviors. The intention to
adopt a specific health behavior such as exercise, dieting, or condom use depends
on, among other things, the perception of one's ability to acquire the necessary
skills to change one's habits. Ralf Schwarzer discusses health behavior theories
that explain and predict behavior change and maintenance. In recent years, the
major theories have been revised by including a self-efficacy factor in the set of
predictor variables because, in many studies, self-efficacy has emerged as the
single best determinant for adopting precaution strategies, abandoning detri-
mental health habits, and preventing relapse. It is not yet clear, however, how this
dimension interacts with perceived risks, outcome expectancies, previous experi-
ence, situational constraints, and other interfering variables. A new model is
presented that may stimulate further inquiry.
Gerjo Kok and his co-workers provide a good example of advanced research
in this field. They studied people's attitudes, perceived social norms, and self-
efficacy toward risk behaviors. They also measured the intention to abstain from
these risk behaviors in the future. It turned out that all three factors predicted
intentions, and that the model would be much less powerful without the self-
efficacy factor. In addition, self-efficacy succeeded in directly influencing actual
behavior. This suggests that low self-efficacy should be conceived as a personal
barrier to the process of adopting appropriate health behaviors. If the individual
has good intentions but feels incompetent to perform the desired action, health
promotion efforts should aim at skill improvement and guided mastery
experience to boost a sense of self-efficacy.
Martin Covington and Carol Omelich have studied the relationship between
the temptation to smoke and the intention to smoke in several thousand high-
school students. They looked at potential mediators that might influence this rela-
tionship. Although a strong direct effect of temptations on intentions emerged,
there were also a number of mediating effects by outcome expectancies and self-
efficacy expectancies. The outcome expectancies were differentiated into affec-
tive consequences and needs consequences, while self-efficacy expectancies were
Preface xiii
distinguished as decision-making strategies, interpersonal strategies, and self-

peer strategies. These cognitions altered the levels of temptation for the students.
Smoking decisions in teenagers, therefore, may be seen as dependent on their
perceived problem-solving capabilities.
The last section deals with applications of self-efficacy theory to physical
symptoms, coping with illness, and rehabilitation of chronic disease. It is
assumed that self-related cognitions exert an effect on illness controllability and
on biological systems that mediate health and illness. The first three chapters of
this section focus on psychological mechanisms in rehabilitation of coronary
heart disease, arthritis, and pulmonary disease.
Craig Ewart investigates the role of self-appraisals in the rehabilitation of
myocardial infarction patients. Convalescence can be prolonged far longer than
desirable by unwarranted fears of a reinfarct when resuming one's normal activi-
ties. An improvement of self-efficacy accelerates progress in exercise and
enhances mood and well-being. Self-efficacy responses to treadmill exercise test-
ing predict activity levels better than medical data. Pretraining self-efficacy levels
predict posttraining gains in terms of perseverance in demanding exercise regi-
mens. It is also of interest that spouses tend to overprotect the patients, but
spouse participation in exercise programs leads to more realistic assessments of
the patients' capabilities.
Halsted Holman and Kate Lorig have launched an arthritis self-management
program to teach patients how to cope with the consequences of their chronic
disease. They found that patients' perceived self-efficacy to cope with their ail-
ment was mediating the outcomes of this program. This was particularly true for
pain and depression. The more self-efficacious the patients had become during
the training, the better they were able to tolerate pain and the less depression was
reported. Beneficial effects were maintained even at a four-year follow-up.
Michelle Toshima, Robert Kaplan and Andrew Ries apply self-efficacy
theory to the rehabilitation of chronic obstructive pulmonary disease and examine
changes in self-appraisals from different sources such as mastery experience and
physiological feedback. The patients reported their self-efficacy expectancies
while undergoing an exercise regimen. Physiological feedback from pulmonary
tests and from treadmill endurance walk tests had a strong influence on self-
efficacy, but mastery experience did not. It is argued that, under certain circum-
stances, information from one source can attenuate information from another.
This may stimulate further research on the joint effects of multiple-source infor-
mation on changes in self-beliefs.
In conclusion, Albert Bandura provides an overview of self-efficacy mecha-
nisms in psychobiological functioning and points to empirical evidence for bio-
chemical effects of self-efficacy in coping with stress, such as autonomous,
catecholamine, and opioid activation. Research on pain control as well as on
immunocompetence demonstrates that people with optimistic self-beliefs are
better off and cope well. They are also at an advantage in the self-management of
xiv Preface
chronic disease and in the rehabilitation process. Self-efficacy improves changes

in health behaviors and thus helps to minimize health risks.
The idea to publish this book this arose when the Freie Universitat Berlin
bestowed upon Albert Bandura an honary doctorate in 1990, at which time a
scientific workshop was organized. The speakers agreed to write up their
presentations, and these papers were supplemented by some invited chapters. I
would like to thank the authors for their excellent contributions. The entire tech-
nical composition of this volume was skillfully accomplished by Mary Wegner.
Her meticulous copy editing and proficient desktop publishing are gratefully
acknowledged.
This book was compiled and edited while I was a visiting scholar at the
University of California at Los Angeles. I appreci.1t~ the generous travel grant
from the Volkswagen Foundation that has made my sojourn possible.
Los Angeles, September 1991 Ralf Schwarzer

I
SELF-EFFICACY AND
HUMAN FUNCTIONING
EXERCISE OF PERSONAL
AGENCY THROUGH THE
SELF -EFFICACY MECHANISM
Albert Bandura
The present chapter analyzes the influential role of perceived self-

efficacy in agent causality. The construction of a sense of personal
efficacy involves a complex process of self-persuasion that relies on
cognitive processing of diverse sources of efficacy information con-
veyed enactively, vicariously, socially, and physiologically. Conver-
gent evidence from diverse lines of research reveals that self-beliefs of
efficacy function as important proximal determinants of human moti-
vation, affect, thought and action. Self-beliefs of efficacy exert their
affects on human functioning through motivational, cognitive and af-
fective intervening processes. Some of these processes, such as affec-
tive states and thinking patterns, are of considerable interest in their
own right, as well as serve as intervening influencers of action. Self-
efficacy beliefs also shape developmental trajectories by influencing
choice of pursuits and selection of environments. Self-efficacy theory
adopts a nondualistic but nonreductional conception of human agency
that operates within a model of triadic reciprocal causation.
The recent years have witnessed a resurgence of interest in self-referent phenom-

ena. One can point to several reasons why self processes have come to pervade
the research in many areas of psychology. Self-generated activities lie at the very
heart of causal processes. They not only give meaning and valence to most ex-
ternal influences, but they function as important proximal determinants of moti-
vation and action. People make causal contributions to their own psychosocial
functioning through mechanisms of personal agency. Among the mechanisms of
agency, the most focal and pervading one involves people's beliefs about their
capabilities to exercise control over events that affect their lives. Self-beliefs of
efficacy influence how people feel, think, and act. This chapter analyzes the
causal function of self-efficacy beliefs and the different psychological processes
through which they exert their effects.
SELF ·EFFICACY CAUSALITY

A central question in any theory of cognitive regulation of motivation, affect
and action concerns the issue of causality. Do self-efficacy beliefs operate as
4 A. Bandura
causal factors in human functioning? This issue has been investigated by a vari-
ety of experimental strategies. Each approach tests the dual-causal link in which
instating conditions affect efficacy beliefs, and efficacy beliefs, in turn, affect
motivation and action. In one strategy, perceived self-efficacy is raised in pho-
bics from virtually non-existent levels to preselected low, moderate, or high
levels by providing them with mastery experiences or simply by modeling coping
strategies for them until the desired level of efficacy was attained (Bandura,
Reese, & Adams, 1982).
As shown in Figure 1, higher levels of perceived self-efficacy are accompa-
nied by higher performance attainments. The efficacy-action relationship is re-
plicated across different dysfunctions and in both intergroup and intrasubject
comparisons, regardless of whether perceived self-efficacy was raised by mastery
experiences or solely by vicarious influence. The vicarious mode of self-efficacy
induction is especially well-suited for demonstrating the causal contribution of
perceived self-efficacy to performance. Individuals simply observe models' per-
formances without executing any actions, make inferences from the modeled in-
formation about their own coping efficacy, and later behave in accordance with
their self-judged efficacy. Microanalysis of efficacy-action congruences reveals a
close fit between perceived self-efficacy and performance on individual tasks.
Another approach to the test of causality is to control, by selection, level of
ability but to vary perceived self-efficacy within each ability level. Collins
(1982) selected children who judged themselves to be of high or low mathemati-
cal efficacy at each of three levels of mathematical ability. They were then given
difficult problems to solve. Within each level of mathematical ability, children
who regarded themselves as efficacious were quicker to discard faulty strategies,
solved more problems (Figure 2), chose to rework more of those they failed, and
did so more accurately than those of equal ability who doubted their efficacy.
Perceived self-efficacy thus exerted a substantial independent effect on perform-
ance. Positive attitude toward mathematics was better predicted by perceived
self-efficacy than by actual ability. As this study shows, people may perform
poorly because they lack the ability, or they have the ability but they lack the
perceived self-efficacy to make optimal use of their skills.
A third approach to causality is to introduce a trivial factor devoid of
information to affect competency, but that can bias self-efficacy judgment. The
impact of the altered perceived self-efficacy on level of motivation is then mea-
sured. Studies of anchoring influences show that arbitrary reference points from
which judgments are adjusted ehher upward or downward can bias the judgments
because the adjustments are usually insufficient. Cervone and Peake (1986) used
arbitrary anchor values to influence self-appraisals of efficacy. Self-appraisals
made from an arbitrary high starting point biased students' perceived self-efficacy
in the positive direction, whereas an arbitrary low starting point lowered students'
appraisals of their efficacy (Figure 3). The initial reference points in a sequence of
performance descriptors similarly biased self-efficacy appraisal (Peake &
Cervone, 1989). In a further study, Cervone (1989) biased self-efficacy appraisal
w w m
() 100 ~100t
Z ~
« « ~.
~ 90 ~ oo~
a: a: o....
a
u. 80 oot- ~
a:
ea: ~
w w :::J
a.. 70 a.. ro~ E.
-I -I
::> ::> >
u. 60 u. 60 t- ~
en ::s
en t'l
en en '<
w w
() 50~ () 501-
() ()
::> INTERGROUP INTRASUBJECT ::> INTERGROUP INTRASUBJECT
en 401:- en 401:-
~
~
~ o~ LOW MEDIUM LOW MEDIUM
LOW MEDIUM HIGH LOW MEDIUM HIGH
LEVEL OF PERCEIVED SELF-EFFICACY
Figure 1 Mean performance attainments as a function of differential levels of perceived self-efficacy. The two left panels
present the relationship for perceived self-efficacy raised by mastery experiences; the two right panels present the
relationship for perceived self-efficacy raised by vicarious experiences. The intergroup panels show the performance
attainments of groups of subjects whose self-percepts of efficacy were raised to different levels; the intrasubject
panels show the performance attainments for the same subjects after their self-percepts of efficacy were successively
VI
raised to different levels (Bandura, Reese, & Adams, 1982).
6 A. Bandura
70
60
1/1 50
z
0
§
.....I
0 40
1/1
w
~
a: 30
:::l
U
U
'"
~ 20
SELF- EFFICACY
10 _HIGH
LOW
0
LOW MEDIUM HIGH
ABILITY LEVEL
Figure 2 Mean levels of mathematical solutions achieved by students as a function of

mathematical ability and perceived mathematical self-efficacy. Plotted from
data of Collins, 1982.
13 30
w
u
~ 12 z
« ~25
u 11
u: iii
:t 10 f- 0:20
w
u..
--J
9
Il.
W
8f- ~ 15
(f)
lL
;:
0 7f-
.....I
)510
W 61-
>
w Sf--
.....I
~ 5
.....I
w
41- --'
LOW NO HIGH LOW NO HIGH

ANCHOR ANCHOR ANCHOR ANCHOR ANCHOR ANCHOR
Figure 3 Mean changes induced in perceived self-efficacy by anchoring influences and

the corresponding effects on level of subsequent perseverant effort (Cervone
& Peake, 1986).
Exercise of Personal Agency 7
by differential cognitive focus on things about the task that might make it
troublesome or tractable. Dwelling on formidable aspects weakened people's be-
lief in their efficacy, but focusing on doable aspects raised self-judgment of capa-
bilities. In all of these experiments, the higher the instated perceived self-
efficacy, the longer individuals persevere on difficult and unsolvable problems
before they quit. Mediational analyses reveal that neither anchoring influences
nor cognitive focus has any effect on motivation when perceived self-efficacy is
partialed out. The effect of the external influences on performance motivation is
thus completely mediated by changes in perceived self-efficacy.
A number of experiments have been conducted in which self-efficacy beliefs
are altered by bogus feedback unrelated to one's actual performance. People
partly judge their capabilities through social comparison. Using this type of effi-
cacy induction procedure, Weinberg, Gould, and Jackson (1979) showed that
physical stamina in competitive situations is mediated by perceived self-efficacy.
They raised the self-efficacy beliefs of one group by telling them that they had
triumphed in a competition of muscular strength. They lowered the self-efficacy
beliefs of another group by telling them that they were outperformed by their
competitor. The higher the illusory beliefs of physical strength, the more physi-
cal endurance subjects displayed during competition on a new task measuring
physical stamina (Figure 4). Failure in a subsequent competition spurred those
with a high sense of self-efficacy to even greater physical effort, whereas failure
further impaired the performance of those whose perceived self-efficacy had
been undermined. Self-beliefs of physical efficacy illusorily heightened in
females and illusorily weakened in males obliterated large preexisting sex
differences in physical strength.
Another variant of social self-appraisal that has also been used to raise or
weaken beliefs of self-efficacy relies on bogus normative comparison. Individ-
uals are led to believe that they performed at the highest or lowest percentile
ranks of the reference group, regardless of their actual performance (Jacobs,
Prentice-Dunn, & Rogers, 1984). Perceived self-efficacy heightened by this
means produced stronger perseverant effort (Figure 5). The regulatory role of
self-belief of efficacy instated by unauthentic normative comparison is replicated
in a markedly different domain of functioning, namely pain tolerance (Litt,
1988). Self-efficacy beliefs were altered by having individuals appear as strong
or weak pain tolerators compared to the capabilities of an ostensibly normative
group. The higher the instated belief in one's capabilities, the greater the pain
tolerance.
Still another approach to the verification of causality employs a contravening
experimental design in which a procedure that can impair functioning is applied,
but in ways that raise perceived self-efficacy. The changes accompanying psy-
chological ministrations may result as much, if not more, from instilling beliefs
of personal efficacy as from the particular skills imparted. If people's beliefs in
their coping efficacy are strengthened, they approach situations more assuredly
and make better use of the skills they have. Holroyd and his colleagues (Holroyd
8 A. Bandura
220
200
U
w
~
:r
I-
<!l 180
z
W
0:
I-
Vl
--'
<!
u 160
iii
>-
:r
c..
140 SELF·EFFICACY
-HIGH
0 - - 0 LOW
0
2 2
COMPETITIVE TRIES
Figure 4 Mean level of physical stamina mobilized in competitive situations as a func-

tion of illusorily instated high or low self-percepts of physical efficacy
(Weinberg, Gould, & Jackson, 1979).
10
20
8
::r:
t5z cO
15
UJ
a: 6 I
Iii w
g
u
z
w 10
u:
Iii
u.. 4 Ui
w a:
w
Ii.. c..
-'
w
en 5
2
LOW HIGH LOW HIGH
SELF·EFFICACY INDUCTION
Figure 5 Mean changes in perceived self-efficacy induced by arbitrary normative com-

parison and the corresponding effects on level of subsequent perseverant
effort (Jacobs, Prentice-Dunn, & Rogers, 1984).
et aI., 1984), demonstrated with sufferers of tension headaches that the benefits of
biofeedback training may stem more from enhancement of perceived coping effi-
cacy than from the muscular exercises themselves. In biofeedback sessions, they
trained one group to become good relaxers. Unbeknownst to another group, they
received feedback signals that they were relaxing whenever they tensed their
muscles. They became good tensers of facial muscles, which, if anything, would
aggravate tension headaches. Regardless of whether people were tensing or
relaxing their musculator, bogus feedback that they were exercising good control
over muscular tension instilled a strong sense of efficacy that they could prevent
the occurrence of headaches in different stressful situations. The higher their
perceived self-efficacy, the fewer headaches they experienced. The actual amount
of change in muscular activity achieved in treatment was unrelated to the
incidence of subsequent headaches.
The findings of the preceding experiments should not be taken to mean that
arbitrary persuasory information is a good way of enhancing self-efficacy beliefs
for the pursuits of everyday life. Rather, these studies have special bearing on
the issue of causality because self-efficacy beliefs are altered independently of a
performance modality and, therefore, cannot be discounted as epiphenomenal by-
products of performance. They demonstrate that changes in self-beliefs of effi-
cacy affect motivation and action. In actual social practice, personal empower-
ment through mastery experiences is the most powerful means of creating a
strong, resilient sense of efficacy (Bandura, 1986, 1988). This is achieved by
equipping people with knowledge, subskills and the strong self-belief of efficacy
needed to use one's skills effectively.
The final way of verifying the causal contribution of self-efficacy beliefs to
human functioning is to test the multivariate relations between relevant determi-
nants and the predicted variable in the theoretical causal model by hierarchical
regression analysis or path analysis. These analytic tools for theory testing indi-
cate how much of the variation in the predicted variable is explained by
perceived self-efficacy when the influence of other determinants is controlled.
The multivariate investigations involve panel designs in which self-efficacy
and the predicted variable are measured on two or more occasions to determine
what effect either factor may have on the other. In some of these studies, per-
ceived self-efficacy is altered by naturally occurring influences during the inter-
vening period. More often, self-efficacy beliefs are altered by experimentally
varied influences. The temporal ordering and systematic variation of perceived
self-efficacy antecedently to the predicted outcome helps to remove ambiguities
about the source and direction of causality. In addition to systematic variation
and temporal priority of the self-efficacy beliefs, controls are applied for poten-
tial confounding variables. The results of such studies reveal that self-efficacy
beliefs usually make substantial contribution to variations in motivation and per-
formance accomplishments (Bandura & Jourden, 1991; Dzewaltowski, 1989;
Locke, Frederick, Lee, & Bobko, 1984; Ozer & Bandura, 1990; Wood &
Bandura, 1989a). The causal contribution of self-efficacy beliefs to
10 A. Bandura
sociocognitive functioning is further documented in comparative tests of the

predictive power of social cognitive theory and alternative conceptual models
(Dzewaltowski, Noble, & Shaw, 1991; Lent, Brown, & Larkin, 1987; McCaul,
O'Neill, & Glasgow, 1988; Siegel, Galassi, & Ware, 1985; Wheeler, 1983).
These diverse causal tests were conducted with different modes of efficacy
induction, diverse populations, both interindividual and intraindividual verifica-
tion, and all sorts of domains of functioning, and with micro level and macro level
relations. The evidence is consistent in showing that perceived self-efficacy
contributes significantly to level of motivation and performance accomplish-
ments. Evidence that divergent procedures produce convergent results adds to
the explanatory and predictive generality of the self-efficacy mediator.
EFFICACY -ACTIV ATED PROCESSES

Self-efficacy beliefs regulate human functioning through four major proces-
ses. They include cognitive, motivational, affective and selection processes.
Some of these efficacy-activated events are of interest in their own right rather
than merely intervening influencers of action. These processes are analyzed in
some detail in the sections that follow.
Cognitive Processes
Self-beliefs of efficacy affect thought patterns that can enhance or undermine

performance. These cognitive effects take various forms. Much human behav-
ior, being purposive, is regulated by forethought embodying cognized goals.
Personal goal setting is influenced by self-appraisal of capabilities. The stronger
the perceived self-efficacy, the higher the goals people set for themselves and the
firmer their commitment to them (Bandura & Wood, 1989; Locke et al., 1984;
Taylor, Locke, Lee, & Gist, 1984). Challenging goals raise the level of motiva-
tion and performance attainments (Locke & Latham, 1990).
Most courses of behavior are initially shaped in thought (Bandura, 1986).
People's beliefs about their efficacy influences the types of anticipatory scenarios
they construct and rehearse. Those who have a high sense of efficacy visualize
success scenarios that provide positive guides for performance. Those who judge
themselves as inefficacious are more inclined to visualize failure scenarios which
undermine performance by dwelling on how things will go wrong. Numerous
studies have shown that cognitive simulations in which individuals visualize
themselves executing activities skillfully enhance subsequent performance
(Bandura, 1986; Corbin, 1972; Feltz & Landers, 1983; Kazdin, 1978). Perceived
self-efficacy and cognitive simulation affect each other bidirectionally. A high
sense of efficacy fosters cognitive constructions of effective actions and cogni-
tive reiteration of efficacious courses of action strengthens self-beliefs of efficacy
(Bandura & Adams, 1977; Kazdin, 1979).
A major function of thought is to enable people to predict the occurrence of

events and to create the means for exercising control over those that affect their
daily lives. Many activities involve inferential judgments about conditional rela-
tions between events. Discovery of such predictive rules requires effective cog-
nitive processing of multidimensional information that contains ambiguities and
uncertainties. The fact that the same predictor may contribute to different effects
and the same effect may have multiple predictors creates uncertainty as to what is
likely to lead to what in probabilistic environments.
In ferreting out predictive rules people must draw on their preexisting knowl-
edge to construct options, to weigh and integrate predictive factors into compos-
ite rules, to test and revise their judgments against the immediate and distal
results of their actions, and to remember which factors they had tested and how
well they had worked. It requires a strong sense of efficacy to remain task ori-
ented in the face of pressing situational demands and judgment failures that can
have important social repercussions.
The powerful influence of self-efficacy beliefs on self-regulatory cognitive
processes is revealed in a program of research on complex organizational
decision-making (Wood & Bandura, 1989b). Much of the research on human
decision-making involves discrete judgments in static environments under non-
taxing conditions (Beach, Barnes, & Christensen-Szalanski, 1986; Hogarth,
1981). Judgments under such circumstances may not provide a sufficient basis
for developing either descriptive or normative models of decision making in
dynamic naturalistic environments which involve repeated judgments in the face
of a wide array of information within a continuing flow of activity under time
constraints and social and self-evaluative consequences. To complicate matters
further, organizational decision making requires working through others and
coordinating, monitoring and managing collective efforts.
The mechanisms and outcomes of organizational decision making do not
lend themselves readily to experimental analysis in actual organizational settings.
Advances in this complex field can be achieved by experimental analyses of
decision making in simulated organizational environments. A simulated environ-
ment permits systematic variation of theoretically relevant factors and precise
assessment of their impact on organizational performance and the psychological
mechanisms through which they achieve their effects.
In this research, executives managed a computer-simulated organization in
which they had to match their supervisees to subfunctions based on their talents
and to learn and implement managerial rules to achieve organizational levels of
performance that were difficult to fulfill. At periodic intervals we measured the
managers' perceived self-efficacy, the goals of group performance they sought to
achieve, the adequacy of their analytic thinking for discovering managerial rules,
and the level of organizational performance they realized.
Social cognitive theory explains psychosocial functioning in terms of triadic
reciprocal causation (Bandura, t 986). In this model of reciprocal determinism,
12 A. Bandura
(1) cognitive, biological and other personal factors, (2) behavior, and (3) environ-
mental events all operate as interacting determinants that influence each other
bidirectionally. Each of the major interactants in the triadic causal structure
----cognitive, behavioral, and environmental-functions as an important constitu-
ent in the dynamic environment. The cognitive determinant is indexed by self-
beliefs of efficacy, personal goal setting, and quality of analytic thinking. The
managerial choices that are actually executed constitute the behavioral determi-
nant. The properties of the organizational environment, the level of challenge it
prescribes, and its responsiveness to managerial interventions represent the envi-
ronmental determinant. Analyses of ongoing processes clarify how the inter-
actional causal structure operates and changes over time.
The interactional causal structure was tested in conjunction with experimen-
tally varied organizational properties and belief systems that can enhance or un-
dermine the operation of self-regulatory determinants. One important belief sys-
tem is concerned with the conception of ability (M. M. Bandura & Dweck, 1988;
Dweck & Leggett, 1988; Nicholls, 1984). Some people regard ability as an
acquirable skill that can be increased by gaining knowledge and perfecting com-
petencies. They adopt a functional learning goal. They seek challenges that pro-
vide opportunities to expand their knowledge and competencies. They regard
errors as a natural part of an acquisition process. One learns from mistakes.
They judge their capabilities more in terms of personal improvement than by
comparison against the achievement of others. For people who view ability as a
more or less fixed capacity, performance level is regarded as diagnostic of inher-
ent cognitive capacities. Errors and deficient performances carry high evaluative
threat. Therefore, they prefer tasks that minimize errors and permit ready display
of intellectual proficiency at the expense of expanding their knowledge and com-
petencies. High effort is also threatening because it presumably reveals low
ability. The successes of others belittle their own perceived ability.
We instilled these different conceptions of ability and then examined their ef-
fects on the self-regulatory mechanisms governing the utilization of skills and
performance accomplishments (Wood & Bandura, 1989a). Managers who
viewed decision-making ability as reflecting basic cognitive aptitude were beset
by increasing self-doubts about their managerial efficacy as they encountered
problems (Figure 6). They became more and more erratic in their analytic think-
ing, they lowered their organizational aspirations, and they achieved progres-
sively less with the organization they were managing. In contrast, construal of
ability as an acquirable skill fostered a highly resilient sense of personal efficacy.
Under this belief system, the managers remained steadfast in their perceived
managerial self-efficacy even when performance standards were difficult to ful-
fill, they continued to set themselves challenging organizational goals, and they
used analytic strategies in efficient ways that aided discovery of optimal
managerial decision rules. Such a self-efficacious orientation paid off in high or-
ganizational attainments. Viewing ability as an inherent capacity similarly lowers
55 22 100
110 20
50 90
18
105
45 16 80
100
Exercise of Personal Agency
14
% SELF-SET GOALS
40 70
95
12 ABILITY CONCEPTION
% ORGANIZATIONAL PERFORMANCE
ACQUIHABLE
INHERENT
EFFICIENT USE OF ANALYTIC STRATEGIES
STRENGTH OF PERCEIVED SELF-EFFICACY

35 90 10 60
1 2 3 1 2 3 1 2 3 1 2 3
TRIAL BLOCKS
Figure 6 Changes in perceived managerial for the organization relative to the preset standard, effective use of analytic strat-
egies, and achieved level of organizational performance across blocks of production trials under conceptions of
ability as an acquirable skill or as an inherent aptitude. Each trial block comprises six different production orders
(Wood & Bandura, 1989a).
13
......
110 ~
70 125
106
120
65
102
3 115
g
60
& 110 98
55
i 94
105
CONTROLLABILITY
50 100 90

HIGH
LOW
1 2 3 1 2 3 1 2 3
TRIAL BLOCKS
?>
Figure 7 Changes in strength of perceived managerial self-efficacy, the performance goals set for the organization, and level t=
of organizational performance for managers who operated under a cognitive set that organizations are controllable
or difficult to control. Each trial block comprises six different production orders (Bandura & Wood, 1989).
l
perceived self-efficacy, retards physical skill development and diminishes

interest in the activity (Jourden, Bandura, & Banfield, 1991).
Another important belief system that affects how efficacy-relevant informa-
tion is cognitively processed is concerned with people's beliefs about the extent
to which their environment is influenceable or controllable. This aspect to the
exercise of control represents the level of system constraints, the opportunity
structures to exercise personal efficacy, and the ease of access to those opportun-
ity structures. Our organizational simulation research underscores the strong
impact of perceived controllability on the self-regulatory factors governing
decision making that can enhance or impede performance (Bandura & Wood,
1989). People who managed the simulated organization under a cognitive set
that organizations are not easily changeable quickly lost faith in their decision-
making capabilities even when performance standards were within easy reach
(Figure 7). They lowered their aspirations. Those who operated under a
cognitive set that organizations are controllable displayed a strong sense of
managerial efficacy. They set themselves increasingly challenging goals and used
good analytic thinking for discovering effective managerial rules. They exhibited
high resiliency of self-efficacy even in the face of numerous difficulties. The
divergent changes in the self-regulatory factors are accompanied by large
differences in organizational attainments.
Path analyses confirm the postulated causal ordering of self-regulatory deter-
minants. When initially faced with managing a complex unfamiliar environment,
people relied heavily on their past performance in judging their efficacy and set-
ting their personal goals. But as they began to form a self-schema concerning
their efficacy through further experience, the performance system is powered
more strongly and intricately by self-perceptions of efficacy (Figure 8). Perceiv-
ed self-efficacy influences performance both directly and through its strong
effects on personal goal setting and proficient analytic thinking. Personal goals,
in tum, enhance performance attainments through the mediation of analytic
strategies.
As previously noted, people judge their capabilities partly through compari-
son with the performances of others. A further experiment in this series examin-
ed how different forms of social comparison affect the mediating self-regulatory
mechanisms and organizational attainments (Bandura & Jourden, 1991). Differ-
ent patterns of performance disparities were conveyed but the findings summar-
ized here are concerned with two that are of special psychological interest. A
progressive mastery pattern showed the managers performing below the com-
parison group at the outset but they gradually closed the gap and eventually sur-
passed their counterparts. A contrasting pattern of progressive decline showed
the managers performing as well as their counterparts at the outset, but then they
began to fall behind and ended well below the comparison group. Figure 9
summarizes the the substantial impact of comparative appraisal on self-regulatory
mechanisms and organizational attainment.
.....
0'1
PERSONAL PERSONAL
GOALS GOALS
)
.23
2) .18
(4
(65
(.4 (.50
)
1)
.26 1)
.62
(.4
.25
.63 (.63) .31 (.34) 39 (.59) .22 (.22) .26 (.42) .28 (.59)
PAST SELF- ANALYTIC SELF- ANALYTIC
PERFORMANCE PERFORMANCE
PERFORMANCE EFFICACY STRATEGIES EFFICACY STRATEGIES
.55 (.79)
.57 (.75) .37 (.55)
Figure 8 Path analysis of causal structures. The initial numbers on the paths of influence are the significant standardized ~
path coefficients; the numbers in parentheses are the first-order correlations. The network of relations on the left t:l:l
half of the figure are for the initial managerial efforts, and those on the right half are for later managerial efforts
(Wood & Bandura, 1989b).
l
I:I1
><
n
(!
0;;.
35 6 90 n
40
0
....
5 d'
50 85 "1
30
'"::l0
60 4 e:.
25 60 >
~
70 3 ::l
<"l
'<
20 75
SO 2
COMPARATIVE PATTERN
LEVEL OF SELF-SATISFACTION
15 90 1 70 Progressive Mastery
Progressive Decline
EFFICIENT USE OF ANALYTIC STRATEGIES

1 2 3 1 2 3 1 2 3 1 2 3
PHASES OF THE EXPERIMENT
Figure 9 Changes in perceived managerial self-efficacy, quality of analytic thinking, and achieved level of organizational
performance across blocks of production orders under comparative appraisal suggesting progressive mastery or pro-
gressive decline relative to a similar comparison group (Bandura & Jourden, 1991).
-....J
18 A. Bandura
Seeing oneself surpassed by similar social referents undermined perceived

self-efficacy, disrupted analytic thinking, created unremitting self-discontent and
increasingly impaired organizational attainments. By contrast, seeing oneself
gain progressive mastery strengthened a sense of personal efficacy, fostered effi-
cient analytic thinking, transformed self-evaluation from self-discontent to self-
satisfaction with accelerating progress and enhanced organizational attainments.
Path analysis conflrms that the different performance trajectories are mediated by
changes in self-regulatory factors.
Motivational Processes
Self-beliefs of efficacy playa central role in the self-regulation of motivation.

Most human motivation is cognitively generated. In cognitive motivation, people
motivate themselves and guide their actions anticipatorily through the exercise of
forethought. They form beliefs about what they can do, they anticipate likely
outcomes of prospective actions, they set goals for themselves and plan courses
of action designed to realize valued futures.
One can distinguish three different forms of cognitive motivators around
which different theories have been built. These include causal attributions, out-
come expectancies, and cognized goals. The corresponding theories are attribu-
tion theory, expectancy-value theory, and goal theory, respectively. Figure 10
summarizes schematically these alternative conceptions of cognitive motivation.
Outcome and goal motivators clearly operate through the anticipation mecha-
nism. Causal reasons conceived retrospectively for prior attainments can also
affect future actions anticipatorily by altering self-appraisal of capability and
perception of task demands.
COGNIZED GOALS
Forelhoughl
ANTICIPATORY
OUTCOME EXPECfANCIES
COGNITIVE PERFORMANCE
MOTIVATORS
PERCEIVED CAUSES OF
Relrospeclive Reasoning SUCCESS AND FAILURE
Figure 10 Schematic representation of conceptions of cognitive motivation based on

cognized goals, outcome expectancies and causal attributions.
The self-efficacy mechanism of personal agency operates in all of these vari-

ant forms of cognitive motivation. Causal attributions and self-efficacy apprais-
als involve bidirectional causation. Self-beliefs of efficacy bias causal attribution
(Alden, 1986; Collins, 1982; Silver, Mitchell, & Gist, 1989). The relative weight
given to information regarding adeptness, effort, task complexity, and situational
circumstances affects self-efficacy appraisal. Causal analyses indicate that the
effects of causal attributions on performance attainments are mediated through
self-efficacy beliefs rather than operate directly on performance (Relich, Debus,
& Walker, 1986; Schunk & Cox, 1986; Schunk & Gunn, 1986; Schunk & Rice,
1986). The stronger the self-efficacy belief, the higher the subsequent
performance attainments.
In expectancy-value theory, strength of motivation is governed jointly by the
expectation that particular actions will produce specified outcomes and the value
placed on those outcomes (Ajzen & Fishbein, 1980; Atkinson, 1964; Feather,
1982; Rotter, 1954). However, people act on their beliefs about what they can
do, as well as their beliefs about the likely outcomes of various actions. The
effects of outcome expectancies on performance motivation are partly governed
by self-beliefs of efficacy. There are many activities which, if done well,
guarantee valued outcomes, but they are not pursued by people who doubt they
can do what it takes to succeed (Beck & Lund, 1981; Betz & Hackett, 1986).
The predictiveness of expectancy-value theory can be enhanced by including the
self-efficacy determinant (Ajzen & Madden, 1986; De Vries, Dijkstra, &
Kuhlman, 1988; McCaul et aI., 1988; Schwarzer, 1992; Wheeler, 1983).
The degree to which outcome expectations contribute independently to per-
formance motivation varies depending on how tightly contingencies between ac-
tions and outcomes are structured, either inherently or socially, in a given domain
of functioning. For many activities, outcomes are determined by level of
accomplishment. Hence, the types of outcomes people anticipate depend largely
on how well they believe they will be able to perform in given situations. In
most social, intellectual, and physical pursuits, those who judge themselves
highly efficacious will expect favorable outcomes, whereas those who expect
poor performances of themselves will conjure up negative outcomes. Thus, in
activities in which outcomes are highly contingent on quality of performance,
self-judged efficacy accounts for most of the variance in expected outcomes.
When variations in perceived self-efficacy are partialed out, the outcomes
expected for given performances do not have much of an independent effect on
behavior (Barling & Abel, 1983; Barling & Beattie, 1983; Godding & Glasgow,
1985; Lee, 1984a, 1984b; Williams & Watson, 1985).
Self-efficacy beliefs account for only part of the variance in expected out-
comes when outcomes are not completely controlled by quality of performance.
This occurs when extraneous factors also affect outcomes, or outcomes are
socially tied to a minimum level of performance so that some variations in
quality of performance above and below the standard do not produce differential
outcomes. And finally, expected outcomes are independent of perceived self-
20 A. Bandura
efficacy when contingencies are discriminatively structured so that no level of

competence can produce desired outcomes. This occurs in pursuits that are rigid-
ly segregated by sex, race, age or some other factor. Under such circumstances,
people in disfavored groups expect poor outcomes however efficacious they
judge themselves to be.
The capacity to exercise self-influence by personal challenge and evaluative
reaction to one's own attainments provides a major cognitive mechanism of moti-
vation and self-directed ness (Bandura, 1991). A large body of evidence is con-
sistent in showing that explicit challenging goals enhance and sustain motivation
(Locke & Latham, 1990). Goals operate largely through self-referent processes
rather than regulate motivation and action directly. Motivation based on aspira-
tional standards involves a cognitive comparison process. By making self-satis-
faction conditional on matching adopted goals, people give direction to their
actions and create self incentives to persist in their efforts until their perform-
ances match their goals. They seek self-satisfactions from fulfilling valued goals
and are prompted to intensify their efforts by discontent with substandard
performances.
Activation of self-evaluation processes through cognitive comparison
requires both comparative factors-a personal standard and knowledge of one's
performance level. Simply adopting a goal, without knowing how one is doing,
or knowing how one is doing in the absence of a goal, has no lasting motivational
impact (Bandura & Cervone, 1983; Becker, 1978; Strang, Lawrence, & Fowler,
1978). But the combined influence of goals with performance feedback
heightens motivation substantially.
Cognitive motivation based on goal intentions is mediated by three types of
self-influences: affective self-evaluative reactions to one's performance, perceiv-
ed self-efficacy for goal attainment, and adjustment of personal standards in light
of one's attainments. Perceived self-efficacy contributes to motivation in several
ways. It is partly on the basis of self-beliefs of efficacy that people choose what
challenges to undertake, how much effort to expend in the endeavor, and how
long to persevere in the face of difficulties (Bandura, 1986, 199]). When faced
with obstacles and failures, people who have self-doubts about their capabilities
slacken their efforts or abort their attempts prematurely and settle for mediocre
solutions, whereas those who have a strong belief in their capabilities exert great-
er effort to master the challenge (Bandura & Cervone, 1983; Cervone & Peake,
1986; Jacobs et aI., 1984; Peake & Cervone, 1989; Weinberg et aI., 1979).
Strong perseverance usually pays off in performance accomplishments.
As previously noted, affective self-reactions provide a dual source of incen-
tive motivation-the anticipated self-satisfaction for personal accomplishment
operates as a positive motivator and discontent with deficient performance func-
tions as a negative motivator. The more self-dissatisfied people are with sub-
standard attainments, the more they heighten their efforts. These two forms of
self-motivators contribute differentially to performance accomplishments
depending on the complexity of the activity. On tasks where success is attainable
solely by increased level of effort, self-discontent with substandard attainments is

the major regulator of performance accomplishments (Bandura & Cervone, 1983,
1986). In contrast, on tasks that make heavy attentional and cognitive demands,
self-satisfaction with personal progress toward challenging standards provides a
positive motivational orientation for performance accomplishments. Strong self-
critical reactions can detract from the intricate task of generating and testing
alternative organizational strategies (Bandura & Jourden, 1991; Cervone, Jiwani,
& Wood, 1991). As people approach or surpass the adopted standard, they set
new goals for themselves that serve as additional motivators. The higher the self-
set challenges, the more effort invested in the endeavor. Thus, notable attain-
ments bring temporary satisfaction, but people who are assured of their capabili-
ties enlist new challenges as personal motivators for further accomplishment.
The contribution of these self-reactive influences to motivation is strikingly
revealed in a study that systematically varied the direction and magnitude of dis-
crepancy between performance and a difficult assigned standard (Bandura &
Cervone, 1986). Inspection of Figure 11 shows that the more sources of self-in-
fluence individuals brought to bear on themselves, the higher the effort they
exerted and sustained to attain what they seek. Taken together this set of self-
reactive influences accounts for the major share of variation in motivation.
to INCREASE IN PERFORMANCE MOTIVATION

140
120
100
BO
60
40
20
0
o
0 1 2
2 3
NUMBER OF SELF-INFLUENCINIl FACTORS
Figure 11 Mean percent change in motivational level as a function of the number of

self-reactive influences operating in given individuals. The three self-
reactive factors included strong perceived self-efficacy for goal attainment;
self-dissatisfaction with substandard performance; and adoption of
challenging standards. Plotted from data of Bandura & Cervone, 1986.
22 A. Bandura
Many theories of motivation and self-regulation are founded on a negative

feedback control model (Carver & Scheier, 1981; Lord & Hanges, 1987; Miller,
Galanter, & Pribram, 1960). This type of system functions as a motivator and
regulator of action through a discrepancy reduction mechanism. Perceived dis-
crepancy between performance and a reference standard motivates action to
reduce the incongruity. Discrepancy reduction clearly plays a central role in any
system of self-regulation. However, in the negative feedback control system, if
performance matches the standard the person does nothing. Such a feedback
control system would produce circular action that leads nowhere. Nor could
people be stirred to action until they received feedback that their performance is
negatively discrepant from the standard.
Self-regulation by negative discrepancy tells only half the story and not
necessarily the more interesting half. People are proactive, aspiring organisms.
Their capacity for forethought enables them to organize and regulate their lives
proactively. Human self-motivation relies on both discrepancy production and
discrepancy reduction (Bandura, 1991). It requires proactive control as well as
reactive control. People motivate and guide their actions through proactive con-
trol by setting themselves valued challenging standards that create a state of dis-
equilibrium and then mobilizing their effort on the basis of anticipatory estima-
tion of what it would take to reach them. Reactive feedback control comes into
play in subsequent adjustments of effort to attain desired results. As previously
shown, after people attain the standard they have been pursuing, those with a
strong sense of efficacy set a higher standard for themselves. Adopting further
challenges creates new motivating discrepancies to be mastered. Similarly, sur-
passing a standard is more likely to raise aspiration than to lower subsequent per-
formance to conform to the surpassed standard. Self-regulation of motivation
and action thus involves a hierarchical dual control process of disequilibrating
discrepancy production followed by equilibrating discrepancy reduction.
There is a growing body of evidence that human attainments and positive
well-being require an optimistic sense of personal efficacy (Bandura, 1986).
This is because ordinary social realities are strewn with difficulties. They are full
of impediments, failures, adversities, setbacks, frustrations, and inequities. Peo-
ple must have a robust sense of personal efficacy to sustain the perseverant effort
needed to succeed. Self-doubts can set in fast after some failures or reverses.
The important matter is not that difficulties arouse self-doubt, which is a natural
immediate reaction, but the speed of recovery of perceived self-efficacy from dif-
ficulties. Some people quickly recover their self-assurance, others lose faith in
their capabilities. Because the acquisition of knowledge and competencies
usually requires sustained effort in the face of difficulties and setbacks, it is
resiliency of self-belief that counts.
In his informative book, titled Rejection, John White (1982) provides vivid
testimony that the striking characteristic of people who have achieved eminence
in their fields is an inextinguishable sense of efficacy and a firm belief in the
worth of what they are doing. This resilient self-belief system enabled them to
override repeated early rejections of their work.
Many of our literary classics brought their authors repeated rejections. The
novelist, Saroyan, accumulated several thousand rejections before he had his flIst
literary piece published. James Joyce's, the Dubliners, was rejected by 22 pub-
lishers. Gertrude Stein continued to submit poems to editors for about 20 years
before one was finally accepted. Now that's invincible self-efficacy. Over a
dozen publishers rejected a manuscript bye. e. cummings. When he finally got it
published by his mother the dedication, printed in upper case, read: With no
thanks to ... followed by the list of 16 publishers who had rejected his offering.
Early rejection is the rule, rather than the exception, in other creative endeav-
ors. The Impressionists had to arrange their own art exhibitions because their
works were routinely rejected by the Paris Salon. Van Gogh sold only one paint-
ing during his life. Rodin was rejected repeatedly by the Ecole des Beaux-Arts.
The musical works of most renowned composers were initially greeted with deri-
sion. Stravinsky was run out of town by an enraged audience and critics when he
flIst served them the Rite of Spring. Many other composers suffered the same
fate, especially in the early phases of their career. The brilliant architect, Frank
Lloyd Wright, was one of the more widely rejected architects during much of his
career.
To turn to more familiar examples, Hollywood initially rejected the incom-
parable Fred Astaire for being only "a balding, skinny actor who can dance a
little." Decca Records turned down a recording contract with the BeatIes with
the nonprophetic evaluation, "We don't like their sound. Groups of guitars are
on their way out." Whoever issued that rejective pronouncement must cringe at
each sight of a guitar. After Decca Records got through rejecting the Beatles,
Columbia Records followed suit with a prompt rejection.
It is not uncommon for authors of scientific classics to experience repeated
initial rejection of their work, often with hostile embellishments if it is too dis-
cordant with what is in vogue at the time. For example, John Garcia, who event-
ually won well-deserved recognition for his fundamental psychological dis-
coveries, was once told by a reviewer of his oft rejected manuscripts that one is
no more likely to find the phenomenon he discovered than bird droppings in a
cuckoo clock. Verbal droppings of this type demand tenacious self-belief to
continue the tortuous search for new Muses. Scientists often reject theories and
technologies that are ahead of their time. Because of the cold reception given to
most innovations, the time between conception and technical realization typically
spans several decades.
The findings of laboratory investigations are in accord with these records of
human triumphs regarding the centrality of the motivational effects of self-beliefs
of efficacy in human attainments. It takes a resilient sense of efficacy to override
the numerous dissuading impediments to significant accomplishments.
24 A. Bandura
It is widely believed that misjudgment breeds dysfunction. The functional

value of veridical self-appraisal depends on the nature of the endeavor. In activi-
ties where the margins of error are narrow and missteps can produce costly or
injurious consequences, personal well-being is best served by highly accurate
self-appraisal. It is a different matter when difficult accomplishments can pro-
duce substantial personal or social benefits and where the personal costs involve
time, effort and expendable resources. Individuals have to decide for themselves
which creative abilities to cultivate, whether to invest their efforts and resources
in endeavors that are difficult to fulfill, and how much hardship they are willing
to endure for pursuits strewn with obstacles.
In most endeavors, optimistic self-appraisals of capability that are not unduly
disparate from what is possible can be advantageous, whereas veridical judg-
ments can be self-limiting. When people err in their self-appraisal they tend to
overestimate their capabilities. This is a benefit rather than a cognitive failing to
be eradicated. If self-efficacy beliefs always reflected only what people can do
routinely, they would rarely fail but they would not mount the extra effort needed
to surpass their ordinary performances. The emerging evidence indicates that the
successful, the innovative, the sociable, the nonanxious, the nondespondent, and
the social reformers take an optimistic view of their personal efficacy to exercise
influence over events that affect their lives (Bandura, 1986). If not unrealistically
exaggerated, such self-beliefs enhance and sustain the level of motivation needed
for personal and social accomplishments. Societies enjoy considerable benefits
from the eventual accomplishments of its persisters.
Affective Processes
The self-efficacy mechanism also plays a pivotal role in the self-regulation of
affective states. One can distinguish three principal ways in which self-efficacy
beliefs affect the nature and intensity of emotional experiences. Such beliefs
create attentional biases and influence how emotive life events are construed and
cognitively represented; they operate in the exercise of control over perturbing
thought patterns; and they sponsor courses of action that transform environments
in ways that alter their emotive potential. These alternative paths of affective in-
fluence are amply documented in the self-regulation of anxiety arousal and
depressive mood.
In social cognitive theory (Bandura, 1986), perceived self-efficacy to exer-
cise control over potentially threatening events plays a central role in anxiety
arousal. Threat is not a fixed property of situational events. Nor does appraisal
of the likelihood of aversive happenings rely solely on reading external signs of
danger or safety. Rather, threat is a relational property concerning the match
between perceived coping capabilities and potentially hurtful aspects of the envi-
ronment. Therefore, to understand people's appraisals of external threats and
their affective reactions to them it is necessary to analyze their judgments of their
coping capabilities which, in large part, determine the subjective perilousness of

environmental events.
People who believe they can exercise control over potential threats do not
conjure up apprehensive cognitions and, hence, are not perturbed by them. But
those who believe they cannot manage potential threats experience high levels of
anxiety arousal. They dwell on their coping deficiencies, view many aspects of
their environment as fraught with danger, magnify the severity of possible threats
and worry about perils that rarely, if ever, happen. Through such inefficacious
thought they distress themselves and constrain and impair their level of function-
ing (Beck, Emery, & Greenberg, 1985; Lazarus & Folkman, 1984;
Meichenbaum, 1977; Sarason, 1975).
That perceived coping efficacy operates as a cognitive mediator of anxiety
and stress reactions has been tested by creating different levels of perceived self-
efficacy and relating them at a micro level to different manifestations of anxiety.
People display little affective arousal while coping with potential threats they
regard with high efficacy. But as they cope with threats for which they distrust
their coping efficacy, their stress mounts, their heart rate accelerates, their blood
pressure rises, and they display increased catecholamine secretion (Bandura et
aI., 1982; Bandura, Taylor, Williams, Mefford, & Barchas, 1985). After per-
ceived efficacy is strengthened to the maximal level by guided mastery, pre-
viously intimidating tasks no longer elicit differential autonomic or
catecholamine reactions.
The foregoing discussion documents how perceived coping self-efficacy af-
fects the neurobiological aspects of emotional states. The types of biochemical
reactions that have been shown to accompany a weak sense of coping efficacy,
such as autonomic and catecholamine activation, are involved in the regulation of
immune systems. Perceived self-inefficacy in exercising control over stressors
also activates endogenous opioid systems (Bandura, Cioffi, Taylor, & Brouillard,
1988). Some of the immunosuppressive effects of inefficacy in controlling stres-
sors are mediated by release of endogenous opioids (Shavit & Martin, 1987)
When opioid mechanisms are blocked by an opiate antagonist, the stress of
uncontrollability loses its immunosuppressive power. These combined findings
identify some of the neurobiological paths through which perceived self-efficacy
can affect immunoregulatory processes.
Several converging lines of evidence show that exposure to stressors without
the ability to control them impairs the immune system (Coe & Levine, 1991;
Maier, Laudenslager, & Ryan, 1985). However, stress activated in the process of
acquiring controlling mastery may have very different effects than stress in aver-
sive situations with no prospect in sight of ever gaining any self-protective effi-
cacy. This view receives some support from examination of immunological
changes accompanying self-efficacy enhancement through guided mastery expe-
riences (Wiedenfeld et al., 1990). The rate with wh ich people acquired a sense of
controlling efficacy was a good predictor of whether exposure to acute stressors
enhanced or suppressed various components of the immune system.
26 A. Bandura
Development of a strong sense of efficacy to control phobic stressors had an

immunoenhancing effect. A slow growth of perceived self-efficacy attenuated
components of the immune system.
Anxiety arousal in situations involving some risks is affected not only by
perceived coping efficacy, but also by perceived efficacy to control distressing
cognitions. The exercise of control over one's own consciousness is summed up
well in the proverb: "You cannot prevent the birds of worry and care from flying
over your head. But you can stop them from building a nest in your head." Per-
ceived self-efficacy in thought control is a key factor in the regulation of
cognitively-generated arousal. It is not the sheer frequency of disturbing cogni-
tions, but the perceived inability to tum them off that is the major source of
distress (Churchill, 1990; Churchill & McMurray, 1990; Kent, 1987; Salkovskis
& Harrison, 1984). Thus, the incidence of aversive cognitions is unrelated to
anxiety level when variations in perceived thought control efficacy are controlled
for, whereas perceived thought control efficacy is strongly related to anxiety
level when extent of frightful cognitions is controlled (Kent & Gibbons, 1987).
The dual regulation of anxiety arousal and behavior by perceived coping effi-
cacy and thought control efficacy is revealed in a study of the mechanisms gov-
erning personal empowerment over pervasive social threats (Ozer & Bandura,
1990). Sexual violence toward women is a prevalent problem. Because any
woman may be a victim, the lives of many women are distressed and constricted
by a sense of inefficacy to cope with the threat of sexual assault. Such concerns
often preoccupy their thinking in situations posing potential risks. To address
this problem at a self-protective level, women participated in a mastery modeling
program in which they perfected the physical skills to defend themselves success-
fully against sexual assailants. Mastery modeling enhanced perceived coping
efficacy and cognitive control efficacy, decreased perceived vulnerability to
assault and reduced the incidence of intrusive aversive thoughts and anxiety
arousal. These changes were accompanied by increased freedom of action and
decreased avoidant social behavior. Path analysis of the causal structure revealed
a dual path of regulation of behavior by perceived self-efficacy: One path was
mediated through the effects of perceived coping self-efficacy on perceived
vulnerability and risk discernment, and the other through the impact of perceived
cognitive control self-efficacy on intrusive aversive thoughts (Figure 12). A
strong sense of coping efficacy rooted in performance capabilities has substantial
impact on perceived self-efficacy to abort the escalation or perseveration of
perturbing cognitions.
Perceived coping efficacy regulates avoidance behavior in risky situations, as
well as anxiety arousal. The stronger the perceived coping self-efficacy the more
venturesome the behavior, regardless of whether self-beliefs of efficacy are
strengthened by mastery experiences, modeling influences, or cognitive simula-
tions (Bandura, 1988). The role of perceived self-efficacy and anxiety arousal in
the causal structure of avoidant behavior has been examined in a number of
studies. The results show that people base their actions on self-beliefs of efficacy
COPING -.67 PERSONAL -.53 RISK PERCEIVED

EFFICACY VULNERABILITY. DISCERNMENT RISK
(.0001) (.001) .34(.04
4)
-.32(.0
49)
.37
.37 (.035) BEHAVIOR
3)
.35(.0 .41 (.01)
COGNITIVE -.44
-.44 NEGATIVE .48
CONTROL ANXIETY
THOUGHTS
EFFICACY (.003) (.004)
Figure 12 Path analysis of the causal structure. The numbers on the paths of influence
are the significant standardized path coefficients; the numbers in parenthe-
ses are the significance levels. The solid line to behavior represents differ-
ent activities pursued outside the home, the hatch line represents avoided
activities because of concern over personal safety (Ozer & Bandura, 1990).
in situations they regard as risky. Williams and his colleagues (Williams,

Dooseman, & Kleifield, 1984; Williams, Kinney, & Falbo, 1989; Williams &
Rappoport, 1983; Williams, Turner, & Peer, 1985) have analyzed by partial cor-
relation numerous data sets from studies in which perceived self-efficacy, antici-
pated anxiety, and phobic behavior were measured. Perceived self-efficacy
accounts for a substantial amount of variance in phobic behavior when
anticipated anxiety is partialed out, whereas the relationship between anticipated
anxiety and phobic behavior essentially disappears when perceived self-efficacy
is partialed out (Table I). Studies of other threatening activities similarly demon-
strate the predictive superiority of perceived self-efficacy over perceived
dangerous outcomes in level of anxiety arousal (Hackett & Betz, 1984; Leland,
1983; McAuley, 1985; Williams & Watson, 1985).
The data taken as a whole indicate that anxiety arousal and avoidant behavior
are largely coeffects of perceived coping inefficacy rather than causally linked.
People avoid potentially threatening situations and activities, not because they
experience anxiety arousal or anticipate they will be anxious, but because they
believe they will be unable to cope successfully with situations they regard as
risky. They take self-protective action regardless of whether or not they happen
to be anxious at the moment. They do not have to conjure up an anxious state be-
fore they can take action. They commonly perform risky activities at lower
strengths of perceived self-efficacy despite high anxiety arousal (Bandura, 1988).
Perceived self-efficacy to exercise control can give rise to despondency as
well as anxiety. The nature of the outcomes over which personal control is
sought operates as an important differentiating factor. People experience anxiety
when they perceive themselves ill equipped to control potentially injurious
events. Attenuation or control of aversive outcomes is central to anxiety. People
are saddened and depressed by their perceived inefficacy in gaining highly
valued outcomes. Irreparable loss or failure to gain valued outcomes figures
28 A. Bandura
Table I
Comparison of the Relation Between Perceived Self-Efficacy and Coping Behavior
When Anticipated Anxiety is Control/ed, and the Relation Between Anticipated Anxiety
and Coping Behavior When Perceived Self-Efficacy is Controlled
Coping Behavior
Anticipated Anxiety Perceived Self-Efficacy

With With
Self-Efficacy Controlled Anticipated Anxiety
Controlled
Williams & Rappoport (1983)

Pretreatment 1 a -.12 .40*
Pretreatment 2 -.28 .S9**
Posttreatment .13 .4S*
Follow-up .06 .4S*
Williams et aI. (1984)
Pretreatment -.36* .22
Posttreatment -.21 .S9***
Williams et aI. (198S)
Pretreatment -.3S* .28*
Posttreatment .OS .72***
Follow-up -.12 .66***
TeIch et aI. (198S)
Pretreatment -.S6*** -.28
Posttreatment .IS .48**
Follow-up -.OS .42*
Kirsch et aI. (1983)
Pretreatment -.34* .S4***
Posttreatment -.48** .48**
Arnow et aI. (198S)
Pretreatment .17 .77***
Posttreatment -.08 .43*
Follow-up -.06 .88**
Williams et aI. (1989)
Midtreatment -.IS .6S***
Posttreatment .02 .47**
Follow-up -.03 .71***
Notes. a The pretreatment phases of some of these experiments include only subjects
selected for severe phobic behavior. They have a uniformly low sense of coping
efficacy. In such instances, the highly restricted range of self-efficacy scores
tends to lower the correlation coefficients in pretreatment phases.
*p < .05, **p < .01, ***p < .001.
prominently in despondency. When the valued outcomes one seeks also protect
against future aversive circumstances, as when failure to secure a job jeopardizes
one's livelihood, perceived self-inefficacy is both distressing and depressing.
Because of the interdependence of outcomes, both anxiety and despair often
accompany perceived personal efficacy.
Several lines of evidence support the role of perceived self-inefficacy in de-
pression. A sense of fulfillment and self-worth can have different sources, each
of which is linked to an aspect of self-efficacy. Perceived self-inefficacy to attain
valued goals that contribute to self-esteem and to secure things that bring satis-
faction to one's life can give rise to bouts of depression (Bandura, 1991; Davis &
Yates, 1982; Kanfer & Zeiss, 1983). A low sense of efficacy to fulfill role
demands that reflect on personal adequacy also contributes to depression
(Cutrona & Troutman, 1986; Olioff & Aboud, 1991).
Self-regulatory theories of motivation and of depression make seemingly
contradictory predictions regarding the effects of negative discrepancies between
attainments and standards invested with self-evaluative significance. Standards
that exceed attainments are said to enhance motivation through goal challenges,
but negative discrepancies are also invoked as activators of despondent mood.
Moreover, when negative discrepancies do have adverse effects, they may give
rise to apathy rather than to despondency. A conceptual scheme is needed that
differentiates the conditions under which negative discrepancies will be moti-
vating, depressing, or induce apathy.
In accord with social cognitive theory, the directional effects of negative goal
discrepancies are predictable from the relationship between perceived self-
efficacy for goal attainment and level of personal goal setting (Bandura &
Abrams, 1986). Whether negative discrepancies are motivating or depressing de-
pends on beliefs on one's efficacy to attain them. Negative disparities give rise
to high motivation and low despondency when people believe they have the effi-
cacy to fulfill difficult standards and continue to strive for them. Negative dis-
parities diminish motivation and generate despondency for people who judge
themselves as inefficacious to attain difficult standards but continue to demand
them of themselves as the basis for self-satisfaction. People who view difficult
goals as beyond their capabilities and abandon them as unrealistic for themselves
become apathetic rather than despondent.
Supportive interpersonal relations can reduce the aversiveness of negative
life events that give rise to stress and depression. However, social support does
more than simply operate as a buffer against stressors. In addition to its protec-
tive function, social support serves a positive proactive function in fostering cop-
ing competencies that alter the threat value of potential stressors. Analyses of
causal structures reveal that perceived interpersonal self-efficacy and social sup-
port contribute bidirectionally to depression. Social support is not a fixed entity
cushioning people against stressors. Rather people have to seek out, cultivate and
maintain social networks. Indeed, the Holahans have shown that people with a
high sense of social efficacy create social supports for themselves. Perceived
30 A. Bandura
social self-efficacy reduces vulnerability to depression both directly and through

the cultivation of socially supportive networks (Holahan & Holahan, 1987a,
1987b). Acquaintances model coping attitudes and strategies, provide incentives
for beneficial courses of behavior, and motivate others by showing that difficul-
ties are surmountable by perseverant effort. Social support enhances perceived
self-efficacy which, in turn, fosters successful adaptation and reduces stress and
depression (Cutrona & Troutman, 1986; Major et al., 1990). A strong sense of
social efficacy thus facilitates development of socially supportive relationships
and social support, in turn, enhances perceived self-efficacy for rendering
adversities less depressogenic.
Much human depression is cognitively generated by dejecting thought pat-
terns. Therefore, perceived self-efficacy to exercise control over ruminative
thought figures prominently in the occurrence, duration, and recurrence of
depressive episodes. Kavanagh and Wilson (1989) found that the weaker the per-
ceived efficacy to terminate ruminative thoughts the higher the depression (r =
-.51), and the stronger the perceived thought control efficacy instilled by
treatment the greater the decline in depression (r = .71) and the lower the
vulnerability to recurrence of depressive episodes (r = -.48). Perceived self-
efficacy retains its predictiveness of improvement and reduced vulnerability to
depressive relapse when level of prior depression is controlled.
The preceding analysis centers on the path of influence from perceived self-
inefficacy to depression. Mood states bias the way in which events are interpret-
ed, cognitively organized, and retrieved from memory (Bower, 1983; Isen, 1987).
Mood and self-efficacy influence each other bidirectionally. Perceived self-
inefficacy breeds depression. Despondent mood diminishes perceived seJf-
efficacy, positive mood enhances it (Kavanagh & Bower, 1985). People then act
in accordance with their mood altered efficacy beliefs, choosing more challeng-
ing activities in a self-efficacious frame of mind than if they doubt their efficacy
(Kavanagh, 1983). Despondency can thus lower self-efficacy beliefs, which
undermine motivation and spawn deficient performances, causing even deeper
despondency. In contrast, by raising perceived self-efficacy that facilitates moti-
vation, aidful cognitive self-guidance and accomplishments, positive mood can
set in motion an affirmative reciprocal process.
Selection Processes
People can exert some influence over their life paths by the environments
they select and the environments they create. Thus far, the discussion has center-
ed on efficacy-related processes that enable people to create beneficial environ-
ments and to exercise control over them. Judgments of personal efficacy also
shape developmental trajectories by influencing selection of activities and envi-
ronments. People tend to avoid activities and situations they believe exceed their
coping capabilities, but they readily undertake challenging activities and pick
social environments they judge themselves capable of handling. Any factor that
influences choice behavior can profoundly affect the direction of personal devel-
opment. This is because the social influences operating in selected environments
continue to promote certain competencies, values, and interests long after the
decisional determinant has rendered its inaugurating effect (Bandura, 1986;
Snyder, 1987). Thus, seemingly inconsequential efficacy determinants of
choices can initiate selective associations that produce major and enduring per-
sonal changes. Selection processes are differentiated from cognitive, motivational
and affective processes because, in prompt dismissal of certain courses of action
on grounds of perceived personal inefficacy, the latter regulative processes never
come into play. It is only after people choose to engage in an activity that they
mobilize their effort, generate possible solutions and strategies of action and
become elated, anxious, or depressed over how they are doing.
The power of self-efficacy beliefs to affect the course of life paths through
choice-related processes is most clearly revealed in studies of career decision-
making and career development (Betz & Hackett, 1986; Lent & Hackett, 1987).
The stronger people's self-belief in their capabilities, the more career options
they consider possible, the greater the interest they show in them, the better they
prepare themselves educationally for different pursuits and the more successful
they are at them. A high sense of decisional self-efficacy is also accompanied by
a high level of exploratory activity designed to aid selection of pursuits (Blustein,
1989).
Biased cultural practices, stereotypic modeling of gender roles, and dissuad-
ing opportunity structures eventually leave their mark on women's beliefs about
their occupational efficacy (Hackett & Betz, 1981). Women are especially prone
to limit their interests and range of career options by self-beliefs that they lack
the necessary capabilities for occupations traditionally dominated by men, even
though they do not differ from men in actual ability. The self-limitation of career
development arises from perceived inefficacy, rather than from actual inability.
By constricting choice behavior that can cultivate interests and competencies,
self-disbeliefs create their own behavioral validation and protection from correc-
tive influence. However, changes in cultural attitudes and practices may be
weakening self-efficacy barriers. Students currently coming through the school
ranks reveal a much smaller disparity between males and females in their beliefs
about their efficacy to pursue successfully different types of careers (Post-
Kammer & Smith, 1985).
Self-efficacy beliefs contribute to the course of social development as well as
occupational pursuits (Perry, Perry & Rasmussen, 1986). The developmental
processes undoubtedly involve bidirectional causation. Beliefs of personal capa-
bilities determine choice of associates and activities, and affiliation patterns, in
tum, affect the direction of self-efficacy development.
Construction of Self-Efficacy as a Self-Persuasion Process

The multiple benefits of a sense of personal efficacy do not arise simply from
the incantation of capability. Saying something should not be confused with
32 A. Bandura
believing it to be so. Simply saying that one is capable is not necessarily self-
convincing, especially when it contradicts preexisting beliefs. For example, no
amount of reiteration that I can fly, will persuade me that I have the efficacy to
get myself airborne. Self-efficacy beliefs are the product of a complex process of
self-persuasion that relies on cognitive processing of diverse sources of efficacy
information conveyed enactively, vicariously, socially, and physiologically
(Bandura, 1986). People cannot persuade themselves of their efficacy if they
regard the information from which they construct their self-beliefs as
unrepresentative, tainted or erroneous.
The cognitive processing of efficacy information involves two separable
functions: The first concerns the types of information people attend to and use as
indicators of personal efficacy. Each of the four modes of conveying information
about personal capabilities has its distinctive set of efficacy indicators. The
second function concerns the combination rules or heuristics people use to weight
and integrate efficacy information from different sources in forming their self-
efficacy beliefs. Self processes govern the construction of such belief systems at
the level of selection, interpretation, and integration of efficacy-relevant
information.
Converging lines of evidence indicate that the self-efficacy mechanism plays
a central role in the exercise of personal agency. The value of a psychological
theory is judged not only by its explanatory and predictive power, but also by its
operative power to enhance the quality of human functioning. Social cognitive
theory provides prescriptive specificity on how to empower people with the
competencies, self-regulatory capabilities and resilient self-belief of efficacy that
enables them to enhance their psychological well-being and accomplishments.
REFERENCES
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behav-
ior. Englewood Cliffs, NJ: Prentice-Hall.
Ajzen, I., & Madden, T. 1. (1986). Prediction of goal-directed behavior: Attitudes,
intentions, and perceived behavioral control. Journal of Experimental Social
Psychology, 22,453-474.
Alden, L. (1986). Self-efficacy and causal attributions for social feedback. Journal of
Research in Personality, 20, 460-473.
Atkinson, J. W. (1964). An introduction to motivation. Princeton, NJ: Van Nostrand.
Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory.
Englewood Cliffs, NJ: Prentice-Hall.
Bandura, A. (1988). Perceived self-efficacy: Exercise of control through self-belief. In
J. P. Dauwalder, M. Perrez, & V. Hobi (Eds.), Annual series of European research in
behavior therapy (Vol. 2, pp. 27-59). Lisse, The Netherlands: Swets & Zeitlinger.
Bandura, A. (1991). Self-regulation of motivation through anticipatory and self-regula-
tory mechanisms. In R. A. Dienstbier (Ed.), Perspectives on motivation: Nebraska
symposium on motivation (Vol. 38, pp. 69-164). Lincoln: University of Nebraska
Press.
Bandura, A., & Abrams, K. (1986). Self-regulatory mechanisms in motivating. apathet-

ic. and despondent reactions to unfulfilled standards. Unpublished manuscript,
Stanford University, Stanford.
Bandura, A., & Adams, N. E. (1977). Analysis of self-efficacy theory of behavioral
change. Cognitive Therapy and Research, 1,287-308.
Bandura, A .• & Cervone, D. (1983). Self-evaluative and self-efficacy mechanisms gov-
erning the motivational effects of goal systems. Journal of Personality and Social
Psychology, 45, 1017- 1028.
Bandura, A., & Cervone, D. (1986). Differential engagement of self-reactive influences
in cognitive motivation. Organizational Behavior and Human Decision Processes,
38,92-113.
Bandura. A., Cioffi, D., Taylor, C. B., & Brouillard. M. E. (1988). Perceived self-effi-
cacy in coping with cognitive stressors and opioid activation. Journal of Personality
and Social Psychology, 55, 479-488.
Bandura. A., & Jourden, F. 1. (1991). Self-regulatory mechanisms governing the im-
pact of social-comparison on complex decision making. Journal of Personality and
Social Psychology, 60. 941-951.
Bandura, A .• Reese, L., & Adams, N. E. (1982). Microanalysis of action and fear arous-
al as a function of differential levels of perceived self-efficacy. Journal of Personality
and Social Psychology, 43,5-21.
Bandura, A., Taylor, C. B., Williams, S. L, Mefford, I. N., & Barchas, 1. D. (1985).
Catecholamine secretion as a function of perceived coping self-efficacy. Journal of
ConSUlting and Clinical Psychology, 53,406-414.
Bandura, A., & Wood, R. E. (1989). Effect of perceived controllability and perform-
ance standards on self-regulation of complex decision-making. Journal of Personality
Bandura, M. M., & Dweck, C. S. (1988). The relationship of conceptions of intelli-
gence and achievement goals to achievement-related cognition. affect and behavior.
Unpublished manuscript.
Barling, J., & Abel, M. (1983). Self-efficacy beliefs and tennis performance. Cognitive
Therapy and Research, 7, 265-272.
Barling, J., & Beattie, R. (1983). Self-efficacy beliefs and sales performance. Journal
of Organizational Behavior Management, 5, 41-51.
Beach, L. R., Bames, V. E., & Christensen-Szalanski, 1. 1. 1. (1986). Beyond heuristics
and biases: A contingency model of jUdgmental forecasting. Journal of Forecasting,
5,143-157.
Beck, A. T., Emery, G., & Greenberg, R. L. (1985). Anxiety disorders and phobias.
New York: Basic Books.
Beck, K. H., & Lund, A. K. (1981). The effects of health threat seriousness and person-
al efficacy upon intentions and behavior. Journal of Applied Social Psychology, II,
401-415.
Becker, L. 1. (1978). Joint effect of feedback and goal setting on performance: A field
study of residential energy conservation. Journal of Applied Psychology, 63, 428-433.
Betz, N. E., & Hackett, G. (1986). Applications of self-efficacy theory to understanding
career choice behavior. Journal of Social and Clinical Psychology, 4,279-289.
Blustein, D. L. (1989). The role of goal instability and career self-efficacy in the career
exploration process. Journal of Vocational Behavior, 35, 194-203.
Bower, G. H., (1983). Affect and cognition. Philosophical Transactions of the Royal
Society of London (Series B), 302,387-402.
34 A. Bandura
Carver, C. S., & Scheier, M. F. (1981). Attention and self-regulation: A control-theory

approach to human behavior. New York: Springer-Verlag.
Cervone, D. (1989). Effects of envisioning future activities on self-efficacy judgments
and motivation: An availability heuristic interpretation. Cognitive Therapy and
Research, 13,247-261.
Cervone, D., Jiwani, N., & Wood, R. (1991). Goal-setting and the differential influence
of self-regulatory processes on complex decision-making performance. Journal of
Personality and Social Psychology, 6/,257-266.
Cervone, D., & Peake, P. K. (1986). Anchoring, efficacy, and action: The influence of
judgmental heuristics on self-efficacy judgments and behavior. Journal of Personality
Churchill, A. C. (1990). Metacognitive self-efficacy and intrusive thought. Unpublished
doctoral dissertation. University of Melbourne, Australia.
Churchill, A. c., & McMurray, N. E. (1990). Self-efficacy and unpleasant intrusive
thought. Submitted for publication.
Coe, C. L., & Levine, S. (1991, in press). Psychoimmunology: An old idea whose time
has come. In P. R. Barchas (Ed.), Social physiology of social relations. Oxford:
Oxford University Press.
Collins, 1. L. (1982, March). Self-efficacy and ability in achievement behavior. Paper
presented at the annual meeting of the American Educational Research Association,
N.Y.
Corbin, C. (1972). Mental practice. In W. Morgan (Ed.), Ergogenic aids and muscular
performance (pp. 93-118). New York: Academic Press.
Cutrona, C. E., & Troutman, B. R. (1986). Social support, infant temperament, and
parenting self-efficacy: A mediational model of postpartum depression. Child
Development, 57, 1507-1518.
Davis, F. W., & Yates, B. T. (1982). Self-efficacy expectancies versus outcome ex-
pectancies as determinants of performance deficits and depressive affect. Cognitive
De Vries, H., Dijkstra, M., & Kuhlman, P. (1988). Self-efficacy: The third factor be-
sides attitude and subjective norm as a predictor of behavioural intentions. Health
Education Research, 3, 273-282.
Dweck, C. S., & Leggett, E. L. (1988). A social-cognitive approach to motivation and
personality. Psychological Review, 95, 256-273.
Dzewaltowski, D. A. (1989). Towards a model of exercise motivation. Journal 0/ Sport
and Exercise Psychology, / /,251-269.
Dzewaltowski, D. A., Noble, J. M., & Shaw, 1. M. (1991). Physical activity participa-
tion: Social cognitive theory versus the theories of reasoned action and planned
behavior. Journal of Sport and Exercise Psychology, 12, 388-405.
Feather, N. T. (Ed.) (1982). Expectations and actions: Expectancy-value models in
psychology. Hillsdale, NJ: Erlbaum.
Feltz, D. L., & Landers, D. M. (1983). Effects of mental practice on motor skill learn-
ing and performance: A meta-analysis. Journal of Sport Psychology, 5, 25-57.
Godding, P. R., & Glasgow, R. E. (1985). Self-efficacy and outcome expectations as
predictors of controlled smoking status. Cognitive Therapy and Research, 9, 583-590.
Hackett, G., & Betz, N. E. (1981). A self-efficacy approach to the career development
of women. Journal o/Vocational Behavior, /8,326-339.
Hackett, G., & Betz, N. E. (1984). Mathematics performance, mathematics self-efficacy,

and the prediction of science-based college majors. Unpublished manuscript, Uni-
versity of California, Santa Barbara.
Hogarth, R. (1981). Beyond discrete biases: Functional and dysfunctional aspects of
judgmental heuristics. Psychological Bulletin, 90, 197-217.
Holahan, C. K., & Holahan, C. J. (1987a). Self-efficacy, social support, and depression
in aging: A longitudinal analysis. Journal of Gerontology, 42,65-68.
Holahan, C. K., & Holahan, C. 1. (1987b). Life stress, hassles, and self-efficacy in
aging: A replication and extension. Journal of Applied Social Psychology, 17, 574-
592.
Holroyd, K. A., Penzien, D. 8., Hursey, K. G., Tobin, D. L., Rogers, L., Holm, 1. E.,
Marcille, P. J., Hall, 1. R., & Chila, A. G. (1984). Change mechanisms in EMG
biofeedback training: Cognitive changes underlying improvements in tension head-
ache. Journal of Consulting and Clinical Psychology, 52, 1039-1053.
Isen, A. M. (1987). Positive affect, cognitive processes, and social behavior. In
L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 20, pp. 203-
253). New York: Academic Press.
Jacobs, B., Prentice-Dunn, S., & Rogers, R. W. (1984). Understanding persistence: An
interface of control theory and self-efficacy theory. Basic and Applied Social Psy-
chology, 5,333-347.
10urden, F. 1., Bandura, A., & Banfield, 1. T. (1991, in press). Impact of conceptions of
ability on self-regulatory mechanisms and skill development. Journal of Sport and
Exercise Psychology.
Kanfer, R., & Zeiss, A. M. (1983). Depression, interpersonal standard-setting, and
judgments of self-efficacy. Journal of Abnormal Psychology, 92, 319-329.
Kavanagh, D. 1. (1983). Mood and self-efficacy. Unpublished doctoral dissertation,
Stanford University, Stanford, CA.
Kavanagh, D. J., & Bower, G. H. (1985). Mood and self-efficacy: Impact of joy and
sadness on perceived capabilities. Cognitive Therapy and Research, 9, 507-525.
Kavanagh, D. 1., & Wilson, P. H. (1989). Prediction of outcome with a group version
of cognitive therapy for depression. Behaviour Research and Therapy, 27,333-343.
Kazdin, A. E. (1978). Covert modeling-Therapeutic application of imagined re-
hearsal. In 1. L. Singer & K. S. Pope (Eds.) The power of human imagination: New
methods in psychotherapy. Emotions, personality, and psychotherapy (pp. 255-278).
New York: Plenum.
Kazdin, A. E. (1979). Imagery elaboration and self-efficacy in the covert modeling
treatment of unassertive behavior. Journal of Consulting and Clinical Psychology,
47, 725-733.
Kent, G. (1987). Self-efficacious control over reported physiological, cognitive and
behavioural symptoms of dental anxiety. Behaviour Research and Therapy, 25,341-
347.
Kent, G., & Gibbons, R. (1987). Self-efficacy and the control of anxious cognitions.
Journal of Behavior Therapy & Experimental Psychiatry, 18, 33-40.
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York:
Springer.
Lee, C. (1984a). Accuracy of efficacy and outcome expectations in predicting perform-
ance in a simulated assertiveness task. Cognitive Therapy and Research, 8, 37-48.
Lee, C. (1984b). Efficacy expectations and outcome expectations as predictors of per-
formance in a snake-handling task. Cognitive Therapy and Research, 8, 509-516.
36 A. Bandura
Leland, E. 1. (1983). Self-efficacy and other variables as they relate to precompetitive

anxiety among male interscholastic basketball players. (Doctoral dissertation, Stan-
ford University, 1983). Dissertation Abstracts International, 44, 1376A.
Lent, R. W., Brown, S. D., & Larkin, K. C. (1987). Comparison of three theoretically
derived variables in predicting career and academic behavior: Self-efficacy, interest
congruence, and consequence thinking. Journal of Counseling Psychology, 34, 293-
298.
Lent, R. W., & Hackett, G. (1987). Career self-efficacy: Empirical status and future di-
rections. Journal of Vocational Behavior, 30,347-382.
Litt, M. D. (1988). Self-efficacy and perceived control: Cognitive mediators of pain
tolerance. Journal of Personality and Social Psychology, 54, 149-160.
Locke, E. A., Frederick, E., Lee, c., & Bobko, P. (1984). Effect of self-efficacy, goals,
and task strategies on task performance. Journal of Applied Psychology, 69, 241-25l.
Locke, E. A., & Latham, G. P. (1990). A theory of goal setting and task performance.
Lord, R. G., & Hanges, P. J. (1987). A control system model of organizational motiva-
tion: Theoretical development and applied implications. Behavioral Science, 32,
161-178.
Maier, S. F., Laudenslager, M. L., & Ryan, S. M. (1985). Stressor controllability, im-
mune function, and endogenous opiates. In F. R. Brush & J. B. Ovennier (Eds.),
Affect, conditioning, and cognition: Essays on the determinants of behavior (pp. 183-
201). Hillsdale, NJ: Erlbaum.
Major, B., Cozzarelli, c., Sciacchitano, A. M., Cooper, M. L., Testa, M., & Mueller,
P. M. (1990). Perceived social support, self-efficacy, and adjustment to abortion.
Journal of Personality and Social Psychology, 59, 452-463.
McAuley, E. (1985). Modeling and self-efficacy: A test of Bandura's model. Journal
of Sport Psychology, 7, 283-295.
McCaul, K. D., O'Neill, K., & Glasgow, R. E. (1988). Predicting the performance of
dental hygiene behaviors: An examination of the Fishbein and Ajzen model and self-
efficacy expectations. Journal of Applied Social Psychology, /8, 114-128.
Meichenbaum, D. H. (1977). Cognitive-behavior modification: An integrative ap-
proach. New York: Plenum Press.
Miller, G. A., Galanter, E., & Pribram, K. H. (1960). Plans and the structure of behav-
ior. New York: Holt.
Nicholls, J. G. (1984). Achievement motivation: Conceptions of ability, subjective ex-
perience, task choice, and performance. Psychological Review, 91, 328-346.
Olioff, M., & Aboud, F. E. (1991). Predicting postpartum dysphoria in primiparous
mothers: Roles of perceived parenting self-efficacy and self-esteem. Journal of
Cognitive Psychotherapy, 5,3-14.
Ozer, E., & Bandura, A. (1990). Mechanisms governing empowerment effects: A self-
efficacy analysis. Journal of Personality and Social Psychology, 58, 472-486.
Peake, P. K., & Cervone, D. (1989). Sequence anchoring and self-efficacy: Primacy
effects in the consideration of possibilities. Social Cognition, 7, 31-50.
Perry, D. G., Perry, L. C., & Rasmussen, P. (1986). Cognitive social learning mediators
of aggression. Child Development, 57, 700-711.
Post-Kammer, P., & Smith, P. (1985). Sex differences in career self-efficacy, conside-
ration, and interests of eighth and ninth graders. Journal of Counseling Psychology,
32,551-559.
Relich, J. D., Debus, R. L., & Walker, R. (1986). The mediation role of attribution and
self-efficacy variables for treatment effects on achievement outcomes. Contemporary
Educational Psychology, J I, 195-216.
Rotter, J. B. (1954). Social learning and clinical psychology. Englewood Cliffs, NJ:
Prentice-Hall.
Salkovskis, P. M., & Harrison, J. (1984). Abnormal and normal obsessions-a replica-
tion. Behaviour Research and Therapy, 22,549-552.
Sarason, I. G. (1975). Anxiety and self-preoccupation. In I. G. Sarason & c. D. Spiel-
berger (Eds.), Stress and anxiety (Vol. 2, pp. 27-44). Washington, DC: Hemisphere.
Schunk, D. H., & Cox, P. D. (1986). Strategy training and attributional feedback with
learning disabled students. Journal of Educational Psychology, 78, 201-209.
Schunk, D. H., & Gunn, T. P. (1986). Self-efficacy and skill development: Influence
of task strategies and attributions. Journal of Educational Research, 79, 238-244.
Schunk, D. H., & Rice, J. M. (1986). Extended attributional feedback: Sequence ef-
fects during remedial reading instruction. Journal of Early Adolescence, 6, 55-66.
Schwarzer, R. (1992). Self-efficacy in the adoption and maintenance of health behav-
iors: A critical analysis of theoretical approaches and a new model. In Schwarzer, R.
(Ed.), Self-efficacy: Thought control of action. New York: Hemisphere (this
volume).
Shavit, Y., & Martin, F. C. (1987). Opiates, stress, and immunity: Animal studies.
Annals of Behavioral Medicine, 9, 11-20.
Siegel, R. G., Galassi, J. P., & Ware, W. B. (1985). A comparison of two models for
predicting mathematics performance: Social learning versus math aptitude-anxiety.
Journal of Counseling Psychology, 32,531-538.
Silver, W. S., Mitchell, T. R., & Gist, M. E. (1989). The impact of self-efficacy on
causal attributions for successful and unsuccessful peiformance. Unpublished manu-
script, University of Washington, Seattle, Washington.
Snyder, M. (1987). Public appearances, private realities: The psychology of self-
monitoring. New York: W. H. Freeman.
Strang, H. R., Lawrence, E. C., & Fowler, P. C. (1978). Effects of assigned goal level
and knowledge of results on arithmetic computation: Laboratory study. Journal of
Applied Psychology, 63, 446-450.
Taylor, M. S., Locke, E. A., Lee, c., & Gist, M. E. (1984). Type A behavior and facul-
ty research productivity: What are the mechanisms? Organizational Behavior and
Human Peiformance, 34,402-418.
Weinberg, R. S., Gould, D., & Jackson, A. (1979). Expectations and performance: An
empirical test of Bandura's self-efficacy theory. Journal of Sport Psychology, J, 320-
331.
Wheeler, K. G. (1983). Comparisons of self-efficacy and expectancy models of occu-
pational preferences for college males and females. Journal of Occupational
Psychology, 56, 73-78.
White, J. (1982). Rejection. Reading, MA: Addison-Wesley.
Wiedenfeld, S. A., O'Leary, A., Bandura, A., Brown, S., Levine, S., & Raska, K.
(1990). Impact of perceived self-efficacy in coping with stressors on components of
the immune system. Journal of Personality and Social Psychology, 59, 1082-1094.
Williams, S. L., Dooseman, G., & Kleifield, E. (1984). Comparative power of guided
mastery and exposure treatments for intractable phobias. Journal of Consulting and
Clinical Psychology, 52,505-518.
38 A. Bandura
Williams, S. L., Kinney, P. J., & Falbo, J. (1989). Generalization of therapeutic changes
in agoraphobia: The role of perceived self-efficacy. Journal of Consulting and Clini-
cal Psychology, 57,436-442.
Williams, S. L., & Rappoport, A. (1983). Cognitive treatment in the natural environ-
ment for agoraphobics. Behavior Therapy, 14,299-313.
Williams, S. L., Turner, S. M., & Peer, D. F. (1985). Guided mastery and performance
desensitization treatments for severe acrophobia. Journal of ConSUlting and Clinical
Williams, S. L., & Watson, N. (1985). Perceived danger and perceived self-efficacy as
cognitive mediators of acrophobic behavior. Behavior Therapy, /6, 136-146.
Wood, R. E., & Bandura, A. (l989a). Impact of conceptions of ability on self-regula-
tory mechanisms and complex decision making. Journal of Personality and Social
Wood, R. E., & Bandura, A. (1989b). Social cognitive theory of organizational
management. Academy of Management Review, 14,361-384.
Author Notes
This chapter includes revised and expanded material from an article published in
The Psychologist as an invited address at the annual meeting of the British
Psychological Society, St. Andrews, Scotland, April 1989.
TWO DIMENSIONS OF PERCEIVED
SELF -EFFICACY: COGNITIVE
CONTROL AND BEHAVIORAL
COPING ABILITY
William J. McCarthy
and Michael D. Newcomb
Confirmatory factor analyses were conducted to test the empirical

justification for distinguishing between perceptions of behavioral cop-
ing ability and perceptions of cognitive control coping ability for
handling environmental challenges. Twenty-four measures of per-
ceived personal effectiveness were collected from 739 young adults.
including measures of perceived ability to have a social impact. asser-
tiveness. leadership style. and dating competence. These items were
submitted to a hierarchical confirmatory factor analysis in a random
half of the sample. As expected. two empirically we/l-justified second
order factors were obtained reflecting perceived cognitive control and
behavioral coping strategies. This factor structure was cross-validated
in the other half of the sample. and separately for males and females.
with all hypothesized features confirmed. Literature on coping strate-
gies, on sex role differences and on self-efficacy predictors is cited as
support for distinguishing between perceived cognitive control and
perceived behavioral coping abilities. Implications of this distinction
for elucidating developmental patterns of drug use and for improving
understanding of relapse in lifestyle change programs are discussed.
Why do individuals rely primarily on intrapsychic coping in some contexts and

behavioral coping in other contexts? Suppose that a middle-aged woman whose
children have grown and left home is distressed to find that her husband pays her
inadequate attention, responding perfunctorily to her comments and showing
more enthusiasm for newspaper reading and TV watching than for talking with
her. Her perceived ability to cope cognitively with this situation may be high or
low. If it is high, then by merely reframing her thoughts, she can palliate or eli-
minate her distress, perhaps by thinking of evidence that she is indeed an inter-
esting person and that her husband pays her no attention because he is preoccu-
pied by his work. If it is low, then the distressful observation that her husband
finds her boring will intrude on her thoughts unless the situation changes.
40 W.1. McCarthy & M. D. Newcomb
Similarly, her perceived ability to cope behaviorally with the situation may
be high or low. If it is high, she may elect to leave him or feel confident that she
can alter his behavior. If it is low, she will feel that the situation is inescapable
and that all of her alternatives are less attractive than the status quo.
Perceptions of personal coping ability have been related to a wide range of
health-related outcomes, including smoking cessation, weight control, alcohol
abuse, exercise, and contraceptive behavior (e.g., Strecher, DeVellis, Becker, &
Rosenstock, 1986, and O'Leary, 1985). As individuals' self-percepts of coping
ability increase, so does the probability of their achieving self-set health goals.
Between the identification of an important self-relevant goal and the ultimate
achievement of the goal are interposed challenges with which the individual must
cope. These challenges may be primarily cognitive or primarily behavioral in
nature. The coping behaviors appropriate for dealing with these challenges have
been termed emotion-focused or problem-focused (e.g., Folkman & Lazarus,
1980). Emotion-focused coping includes such behaviors as avoidance, intellec-
tualization, isolation, suppression, and magical thinking. Problem-focused cop-
ing includes such behaviors as information-seeking, cognitive problem-solving,
inhibition of action and direct action. The perception that one can effectively
implement emotion-focused or problem-focused coping can be termed perceived
cognitive control ability and perceived behavioral coping ability, respectively.
Other literature on coping has promoted a distinction between behavioral and
cognitive ways of coping. For example, Pearlin and Schooler (1978) discussed
three major categories of coping responses, two of which involved cognitive
strategies to reduce or eliminate stress, whereas the third concerned the active
manipulation of the environment. In her review, Taylor (1986) identified four
types of control that mediated the effects of coping with stressors, but concluded
that these four types of control could be reduced to two: (a) changing thoughts
with respect to the stressor, and (b) taking some action with respect to the stres-
sor. If people differ in whether they rely primarily on cognitive or behavioral
means of coping with a challenge, they also probably differ in their perceived
ability to use either cognitive control or behavioral strategies for coping with the
challenges. The following report seeks to confirm the validity and usefulness of
distinguishing between perceived cognitive control and behavioral coping ability
through confirmatory factor analysis of young adult data on coping strategies and
through example.
Although we find it useful to distinguish between perceived cognitive control
ability and perceived behavioral coping ability, we note that efficacious behavior
is rarely a function exclusively of only one of these. Characteristics of the con-
text (such as the amount of freedom individuals have to change the environment)
and characteristics of the individual (such as age) determine which type of per-
ceived coping ability is the more important contributor to self-perceived ability to
perform the desired behavior.
Cognitive/Behavioral Efficacy 41
In their study of adult responses to 1,332 stress episodes, Folkman and

Lazarus (1980) noted that both problem-focused and emotion-focused coping
were used to cope with 98% of the episodes. They also noted that the importance
of the type of coping varied, however, with context and according to the charac-
teristics of the individuals. Cognitive strategies were employed most frequently
in situations where the individual was relatively helpless to bring about the
desired behavior by themselves, such as when recovering from an illness.
Problem-focused strategies were employed more frequently in work situations.
Folkman and Lazarus found that the importance of type of coping varied with
gender, with men relying more heavily than women on problem-focused coping
even when the context permitted only emotion-focused coping. It should be not-
ed, again, that distinguishing conceptually between perceived cognitive control
and perceived behavioral coping ability should not imply that we view these con-
cepts as independent. Folkman and Lazarus (1980), in fact, observed a mean cor-
relation of .44 between emotion-focused and problem-focused coping across
three different samples.
This study was designed to confirm the reasonableness of distinguishing
generically between perceived cognitive coping ability and perceived behavioral
coping ability as separate components of self-efficacy. The importance of this
distinction, if accepted, is that it would be likely to stimulate more careful exami-
nation of the contextual, temporal, and individual influences on percepts of self-
efficacy. Individuals may have similar perceptions concerning their respective
abilities to accomplish their jobs or to effect major lifestyle changes, but never-
theless vary greatly in their perceived ability to cope with the behavioral or cog-
nitive challenges associated with accomplishing the desired behavior. Moreover,
within individuals the relative importance of perceived cognitive control and be-
havioral coping ability may vary with time, age and the individual's experience
with coping with the specific challenge.
In settings such as prisons, where inmates would find that a problem-focused
coping strategy is often inappropriate, individuals can still vary in their ability to
cope with environmental stressors, depending on their perceived ability to regu-
late their thoughts. In the same vein, children's responses to environmental chal-
lenges often are limited to cognitive coping strategies because their dependency
on adults and their immaturity are such as to obviate the use of behavioral coping
strategies. A child who is sexually abused by an adult relative, for instance, typi-
cally relies on cognitive strategies to cope with the situation, especially disasso-
ciation, and later, amnesia (Courtois, 1988).
Contrariwise, the theoretical necessity for the separate concept of a perceived
ability to cope behaviorally seems justified by attempts to relate the use of coping
strategies to spontaneous major lifestyle change. Spontaneous major lifestyle
changes, such as taking up jogging, reducing the percentage of calories in one's
diet derived from fat, and adopting a child, put a premium on behavioral coping
relative to cognitive control coping, because there are simply too many changes
in day-to-day behaviors resulting from the major lifestyle change in question to
be anticipated cognitively. It is reasonable to assume that persons with strong

beliefs in their general ability to change their social and physical environment
will be more likely to embark on major voluntary lifestyle changes than are per-
sons who perceive themselves as generally having weak behavioral coping stra-
tegies. When discussing the influences on adoption of major lifestyle change one
has to speak of "general" ability because many of the specific challenges that
follow the major life style change are unanticipated at the time of the decision.
Successful voluntary lifestyle change is often accompanied by multiple changes
in how the individual interacts with her/his social and physical environment. A
successful change in diet, for instance, will typically be accompanied by changes
in shopping habits, changes in cooking habits, and changes in where, when and
with whom to dine out. For predicting the success of major lifestyle change ef-
forts, one's ability to cope intrapsychically with specific challenges seems less
relevant than one's general ability to have a behavioral impact on one's environ-
ment. This focus on "general" ability might seem inappropriate for a discussion
of self-efficacy as it is classically defined (Bandura, 1977). This focus is consis-
tent, however, with recent demonstrations that global judgments about a subject's
ability to train employees or to influence organizational performance can influ-
ence their perception of their own organizational ability (Bandura & Wood,
1989; Wood & Bandura, 1989) and subsequent organizational performance.
Description of Proposed Study
For this study, we followed the strategy of Ryckman, Thornton, and Cantrell
(1982). A comprehensive range of 24 measures of personal effectiveness was
administered to a community sample of young adults being followed in a longi-
tudinal study of growth and development. These measures were submitted to a
hierarchical, confirmatory factor analysis in a random half of the sample and
cross-validated in the other half. Based on the literature discussed above, we hy-
pothesized finding two second order factors of perceived behavioral ability
(Perceived Behavioral Coping Ability) and perceived cognitive control ability to
cope with environmental challenges (Perceived Cognitive Control Ability).
Multiple assessments of personal effectiveness included measures of perceived
ability to have a social impact, general assertiveness, dating competence, social
support, depression, perceived loss of control, purpose in life, and leadership
style. These were selected to provide a comprehensive range of measures of the
subjects' cognitive and behavioral skills and emotional states related to interper-
sonal relations. We expected that Perceived Behavioral Coping Ability would be
reflected in first-order factors of social impact efficacy (perceived ability to have
a social impact), general assertiveness, social resources, dating competence, and
leadership style because these represent ways of behaviorally operating on the
environment. On the other hand, we expected that Cognitive Control Coping
Ability would be reflected in constructs of social impact efficacy, depression,
purpose in life, and perceived loss of control because these involve internal cop-
ing or cognitive qualities. Expected social impact efficacy was expected to load
CognitivelBehavioral Efficacy 43
on both second-order factors because this measure included cogmtIve and

behavioral coping items (Blatt, Quinlan, Chevron, McDonald, & Zuroff, 1982).
Table 1
Description of Sample
Characteristic N %
Sex
Male 221 30
Female 518 70
Age
Mean 21.93
Range 19 - 24
Ethnicity
Black 111 15
Hispanic 72 10
White 490 66
Asian 66 9
High School Graduate
Yes 684 93
No 55 7
Living Situation
Alone 28 4
Parents 343 46
Spouse 77 10
Spouse and Child 56 8
Cohabitation 67 9
Dormitory 40 5
Roommates 96 13
Other 32 4
Number of Children
None 619 84
One 106 14
Two or more 14 2
Current Life Activity
Military 23 3
Junior College 87 12
Four-year college 153 21
Part-time job 102 14
Full-time job 343 46
Other 31 4
Income for Past Year
None 71 10
Under $5,000 242 33
$5,001 to $15,000 334 45
Over $15,001 92 12
44 w. J. McCarthy & M. D. Newcomb
The analyses proceeded in steps, first testing for sex differences on the
measures. Then the adequacy of the hypothesized latent measurement model was
tested followed by testing the second-order factor model of the primary-order
latent variables. These analyses were conducted in the derivation sample, and
then the second-order factor results were confirmed in the cross-validation
sample, and in separate samples of men and women.
METHOD
SUbjects
Participants in this study were 739 young adults who completed an eight-year
(fifth wave data assessment point) longitudinal study of adolescent and young
adult development. Data were collected initially from 1,634 students in the
seventh, eighth, and ninth grades at II randomly-selected Los Angeles County
schools. At this young adult follow-up, each participant was paid $12.50 to
complete the questionnaire and all subjects were apprised of a grant of confiden-
tiality given by the U.S. Department of Justice. Forty-five percent of the original
sample participated as young adults. The loss of subjects due to attrition over
the eight years has been shown not likely to bias the results adversely (Newcomb,
1986; Newcomb & Bentler, 1988a).
Table I presents a description of the sample. As evident, 30% were men and
70% were women, which parallels the sex distribution in the initial sample and
does not reflect differential attrition by sex. Most were employed full-time and
represented varied ethnic backgrounds. Additional information is provided about
their current living arrangements, income, and current life pursuits. When these
sample characteristics are compared with national samples of young adults (e.g.,
Bachman, O'Malley, & Johnston, 1984; Miller et aI., 1983) or other studies of
young adult populations (e.g., Donovan, Jessor, & Jessor, 1983; Kandel, 1984)
very similar patterns emerged. Consequently, we consider this sample to be
reasonably representative of young adults in general.
Measures
Table 2 presents a listing of the 24 variables used in this study. They are
organized according to the latent construct they are hypothesized to reflect. For
instance, the latent construct of Social Impact Efficacy is assumed to generate the
variation in three observed indicators called inner resources, independence, and
others' respect. For factors which tend to be unidimensional in nature, three
measured-variable indicators were constructed from the items to reflect the latent
factor or construct. This was done, since, as a rule, it is recommended to have at
least three, highly correlated indicators to identify a latent construct (e.g., Bentler
& Newcomb, 1986). This was done on the self-efficacy, dating competence,
general assertiveness, and purpose of life factors. Standard univariate statistics
for each variable are also given in the table. Below we describe how each
variable was assessed in regard to the latent construct it represents.
Social impact efficacy. Three scales are used to reflect the Social Impact
Efficacy construct. These were derived from the five-item scale of efficacy
developed by Blatt et al. (1982). Responses to these five items were given on a
five-point anchored rating scale that ranged from strongly disagree (1) to
strongly agree (5). The items were factor analyzed and found to reflect a unitary
construct (only one eigenvalue greater than 1.00 and all factor loadings were
greater than .4 on the first unrotated factor). As a result, these five items were
combined into three scales based on content. Inner resources was assessed with a
single item-"I have many inner resources." Independence was the average of
two items-"I am a very independent person" and "I set my personal goals as
high as possible." Others' respect was the average of two items-"Others have
high expectations of me" and "What 1 do and say has a great impact on those
around me."
General assertiveness and dating competence. Dating Competence was as-
sessed by three scales (dating I, dating 2, and dating 3) derived from a nine-item
scale of social competence in dating situations. General Assertiveness was asses-
sed by three scales (assertive I, assertive 2, and assertive 3) obtained from a nine-
item social assertiveness scale. The total Dating Competence and Social Asser-
tiveness scales were developed by Levenson and Gottman (1978), and in several
studies had quite good discriminant validity in both normal and clinic samples.
In the derivation study the entire Dating Competence scale had an internal con-
sistency reliability alpha of .92, while for the General Assertiveness scale the
alpha was .85. Latent constructs derived from these scales have also been used in
a study of sexual behavior and responsiveness (Newcomb, 1984).
Social resources. Three questions were asked to determine the quantity or
amount of social supports as perceived in three life contexts. The first item asked
"How many clubs, groups, or organizations do you belong to (including church
groups)?" The second item asked "How many friends do you reaJJy feel close
to?" And the third item asked "How many family members or relatives can you
talk to about things personal to you?" Responses were given on a rating scale
that ranged from none to nine or more. These items were specifically developed
for this research project, but are similar to standard measures of social support
that focus on amount of social resources and correlate quite highly with satisfac-
tion with social support from various types of social networks (Newcomb &
Bentler, 1988b).
Depression. The 20-item depression scale from the Center for Epidemio-
logic Study of Depression (CES-D) was completed by all subjects. The devel-
opment, validities, and reliabilities of the measure have been reported elsewhere
(Husaini, Neff, Harrington, Hughes, & Stone, 1980; Radloff, 1977; Weissman,
Sholomskas, Pottenger, Prusoff, & Locke, 1977). Participants were asked to rate
their frequency of occurrence for each of the 20 symptom items during the past
week on a scale from none (0) to 5 -7 days (3). The 20 items were factor
analyzed in this sample and found to contain four distinct factors, which is con-
sistent with previous attempts to determine the factor structure of the CES-D
(e.g., Clark, Aneshensel, Frerichs, & Morgan, 1981; Radloff, 1977; Roberts,
1980). The four factors included positive affect, negative affect, impaired moti-
vation, and impaired relationships. Items were averaged into the respective four
factors and were used as indicators of a general latent construct of Depression.
Perceived loss of control. Three single-item variables are hypothesized to
reflect the construct of Perceived Loss of Control. Subjects were asked to rate
their degree of agreement with three statements: (I) "I feel I am not in control of
my life," (2) "} feel that whether or not I am successful is just a matter of luck
and chance, rather than my own doing," and (3) "} feel that others are running my
life for me." Responses were given on a seven-point anchored rating scale that
ranged from strongly disagree (I) to strongly agree (7). Cronbach's alpha for
these three items was .65. This construct assesses a general lack of control over
life events, and has been validated in other samples and studies (Newcomb, 1986;
Newcomb & Harlow, 1986).
Purpose in life. The Purpose in Life test (Crumbaugh, 1968; Crumbaugh &
Maholic, 1964, 1969) consists of 20 items designed to assess one's level of or
purpose in life. Each item was rated on a seven-point anchored rating scale rang-
ing from strongly disagree (I) to strongly agree (7). Previous research on the
Purpose in Life test indicated that it contains several small primary factors and
one large general factor (Harlow, Newcomb, & Bentler, 1987). For purposes of
this study, the 20 items were randomly assigned into three scales (PIL I, PIL 2,
and PIL 3) which were used as manifest indicators of a latent construct of
Purpose in Life.
Leadership style. Two personality scales, ambition and leadership, were
used to reflect the construct of Leadership Style. These traits were assessed
using a self-rating test modified for this research program, but based on the
Bentler Psychological Inventory (BPI; Bentler & Newcomb, 1978; Huba &
Bentler, 1982). Although the BPI was developed with multivariate methods, the
items have a high degree of face validity. Half of the items for each trait are
reverse-scored to minimize response bias or acquiescence. Four items were used
to assess each trait and each item was rated on a 5-point bipolar scale. Thus,
each scale had a range of 4 to 20. The BPI has proved useful in studies of marital
success and failure (Bentler & Newcomb, 1978), criminal behavior (Huba &
Bentler, 1983), and adolescent substance use (Huba & Bentler, 1982). The
period-free test-retest reliability for ambition was .72 and the reliability for
leadership was .71 (Stein, Newcomb, & Bentler, 1986).
Analyses
Our first set of analyses use point-biserial correlations to test for mean differ-
ences between men and women on each of the 24 variables. We next use a con-
firmatory factor analysis with latent variables to evaluate the adequacy of the
hypothesized factor structure (e.g., Bentler, 1980; Bentler & Newcomb, 1986), in
a random half of the sample, which we call the derivation sample. An inspection
of the skew and kurtosis estimates for the 24 observed measures indicates that
they are relatively normally distributed. As a result we will use the maximum
likelihood structural model estimator, which requires multivariately normal data
(e.g., Bentler, 1983, 1986). If the initial hypothesized model does not adequately
reflect the data (which is common in models with many variables and many sub-
jects), we will modify the model until an acceptable fit is achieved, in a manner
which will not disturb the critical features of the model. These empirical model
modifications will be guided by the multivariate Lagrangian Multiplier test for
adding parameters and the multivariate Wald test for deleting parameters (Bentler
& Chou, 1986). Once this is accomplished, we will attempt to confirm our two
hypothesized second-order constructs in this model, making modifications where
necessary. This final model will be tested separately in the other random half of
our sample (called the cross-validation sample), as well as the samples of men
and women to determine whether the second-order factor structure is an accurate
representation in these samples. The final model will also be tested in the origi-
nal derivation sample, but without the empirical modifications to establish
whether the model modifications may have distorted or biased the final results.
RESULTS
Sex Differences
Mean differences between men and women on the 24 variables were tested
using point-biserial correlations. Males were coded I and females were coded 2,
so that a positive correlation indicates that the women had the larger value and a
negative correlation indicates that the men had the larger value. These mean dif-
ference correlations are presented in the right-hand column of Table 2.
Of the 24 variables, significant mean differences were found on 14 of them.
These differences indicate that the women, compared to the men, felt that they
had fewer inner resources, less independence, less respect from others, less as-
sertiveness (on all three scales), a smaller number of friends they could rely on,
less positive affect, more negative affect, more impaired motivation, others con-
trolled her life more, slightly less dating competence (only one scale significantly
different), less ambition, and fewer leadership qualities. Although there were
many mean differences between the men and women, the magnitude of these dif-
ferences was quite small. For instance, the largest difference accounted for only
four percent of the variance between groups (on ambition). Based on these rather
small in magnitude mean differences between men and women. and previous re-
sults indicating that there were not different factor structures for men and women
on social support and loneliness variables (Newcomb & Bentler, 1986) and on
physical health status indicators (Newcomb & Bentler, 1987), we will collapse
across sex for the bulk of the remaining analyses. However, we will test our final
48 W. 1. McCarthy & M. D. Newcomb
model in the separate samples of men and women to determine whether we may
have obscured any important findings by combining the men with the women.
Table 2
Summary o/Variable Characteristics and Sex Mean Difference Tests
Sex
FactorNariable Mean Range SD Skew Kurtosis Difference
a
rpb
Social Impact Efficacy

Inner resources 3.87 1-5 .75 -.55 .61 -.18***
Independence 3.86 1-5 .76 -.76 .41 -.11 **
Others respect 3.64 1-5 .62 -.39 .56 -.08*
General Assertiveness
Assertive 1 8.89 4-13 1.84 -.13 -.38 -.13**
Assertive 2 9.07 4-14 1.90 .03 -.32 -.18***
Assertive 3 8.74 4-14 1.87 .05 -.24 -.09*
Social Resources
Number of family
members 3.48 0-9 2.44 .91 .12 .02
Number of friends 4.04 0-9 2.34 .71 -.10 -.14***
Number of organizations 1.34 0-9 1.36 1.21 2.54 -.06
Depression (CES-D)
Positive affect 2.34 0-3 .64 -.96 .32 -.10**
Negative affect .63 0-3 .64 1.20 1.20 .15***
Impaired motivation .71 0-2.75 .46 .79 .79 .08*
Impaired relationships .34 0-3 .47 1.72 3.30 .00
Purpose in Life
PIL I 5.57 2-7 1.80 -.81 .22 -.04
PIL2 5.30 2.4-7 1.80 -.58 .06 -.06
PIL 3 5.41 2-7 1.77 -.57 .13 -.05
Perceived Loss of Control
Not in control 2.26 1-7 1.51 1.31 1.05 -.02
Powerless 2.28 1-7 1.45 1.20 .82 .03
Others control life 2.73 1-7 1.50 .87 .01 .11 **
Dating Competence
Dating 1 8.84 4-13 1.75 -.06 -.28 -.08*
Dating 2 9.25 3-13 1.82 -.19 -.23 -.03
Dating 3 9.96 4-14 1.87 -.14 -.17 -.05
Leadership Style
Ambition 14.52 4-20 3.53 -.41 -.49 -.20***
Leadership 14.29 5-20 2.83 -.13 -.30 -.16***
Note. a Males were coded I and females were coded 2, so that a positive point-biserial
correlation indicates that the females had the larger value.
*p < .05; **p < .01; ***p < .001.
First-Order Latent Factor Model

This first sequence of models are tested on a random half (n = 370) of our
total sample, which we call our derivation sample. In the initial confirmatory fac-
tor model, the eight latent constructs were hypothesized to "cause" or generate
the variation in the 24 observed variables. The factor structure of this first model
was "pure" in that each observed variable was allowed to load on only one latent
construct. For instance, inner resources was assumed to be an indicator only of
Social Impact Efficacy. This assumption of mutual exclusivity may be an overly
constrained imposition on the model, since many of the variables are concep-
tually similar and may in fact reflect more than one underlying quality.
To identify the model all factor loadings were freed, the variances of the con-
structs were fixed at unity, and all factor intercorrelations were allowed to be
freely estimated. This initial confirmatory factor model did not fit the data to an
acceptable degree, chi2 (df = 224, N = 370) = 480.17, p < .001, NR (normed fit
index: Bentler & Bonett, 1980) = .86. Latent-factor intercorrelations for this mo-
del are presented in the upper triangle of Table 3. All hypothesized factor load-
ings were highly significant, p < .00]. The NFl was sufficiently large to suggest
that an acceptable model could be achieved by adding several small empirically
determined parameters that were not hypothesized in the initial model.
Based on an examination of selected modification indices for additional
factor loadings and correlated uniquenesses (Bentler & Chou, 1986), five non-hy-
pothesized factor loadings and 22 correlations among manifest variable residuals
were added to the model. With these additions the mode] adequately fit the data,
chi 2 (df = 196, N = 370) = 181.97, p = .76, NFl = .95. This new model was a sig-
nificant improvement over the initial model (p < .(01). A summary of all model
fit statistics and difference chi2 tests are given in Table 4. To test whether the ad-
dition of these empirically-determined parameters distorted the substantive inter-
pretation of the model, the latent-factor intercorrelations from the initial model
were correlated with those obtained in the final, modified model. These param-
eters were correlated greater than .99. As a result, the final model was not
considered biased due to the model modifications.
Standardized factor loadings and residual variances for the final first-order
confirmatory factor model are given in Figure I. The rectangles represent the ob-
served variables, the large circles indicate the latent constructs, and the small cir-
cles reflect residual variances of the observed variables. These five non-hypothe-
sized factor loadings tend to be small in magnitude (only one is over .40) and all
are in interpretable directions. For instance, general assertiveness also negatively
influenced the perception that others control your life. The latent-factor intercor-
relations for this model are given in the lower triangle of Table 3. The absolute
value of the correlations ranged from a low of .06 to a high of .88, and II were
higher than .5. In other words, many of the constructs appear to be highly
correlated and may reflect higher-order factors, as originally hypothesized.
50 W. J. McCarthy & M. D. Newcomb
Figure 1 Final confirmatory factor analysis model for the derivation sample. Large circles represent latent factors, rectangles
observed variables, and small circles residuals. Not depicted in the figure for reasons of clarity are two-headed
arrows (correlations) joining all possible pairs of latent factors. Parameter estimates are standardized and residual
.64 Inner resources .60
.60 Social
.64 Independence Impact
.74 Efficacy
.66 Others respect
variables are variances. Significance levels were determined by critical ratios (Up < .01; ***p < .001).
.38 Assertive 1 .79
.74 Gemeral
.45 Assertive 2 Assertive-
.59 ness
.66 Assertive 3
-.28
.67 Number of
familv members .57
Number of .57 Social
,67
friends Resources
.92 Number of
organizations .57
.47 Positive affect

.-.40
.09
.22 Negative affect .88
Depression
Impaired .76 (CES-D)
.42'
motivation
Impaired .48
.59,
relationships .26
PIL1 .43
.37 .78
.39 PIL2 .83 Purpose

in Life
.85
.31. PIL3
.50 -.35
Not in control 71
Perceived
.74 Powerless .51
Loss of
.51 Control
.58 Others control
ffe
.21 Dating 1
.89
.79 Dating
.37, Dating 2
Competence
.72
.48 Dating 3
Ambition ,63
.59, Leader-
.73 ship
.44 Leadership Style
Table 3
Factor Intercorrelations Between the Initial (Upper Triangle) and Final (Lower
Triangle) First-Order Confirmatory Factor Models
Factor I II III IV V VI VII VIII
I Social Impact Efficacy 1.00 .65 .44 -.44 .71 -.59 .51 .62
II General Assertiveness .60 1.00 .30 -.31 .50 -040 .58 .67
III Social Resources .40 .31 1.00 -.37 .50 -.30 048 .20
IV Depression (CES-D) -.44 -.34 -.29 1.00 -.59 .73 -.35 -.22
V Purpose in Life .78 .46 .50 -.53 1.00 -.88 .54 .30
VI Perceived Loss of
Control -.61 -.39 -.25 .61 -.84 1.00 -.48 -.32
VII Dating Competence .56 .61 049 -.32 .52 -A 1.00 .48
VIII Leadership Style .61 .56 .12 -.06 .27 -.25 AI 1.00
Note. r between initial and final correlations> .99. All correlations are significant at
p < .01.
Second-Order Confirmatory Factor Models

Based upon our theoretical position, which hypothesized that two second-
order factors should underlie the construct of self-efficacy, two second-order
factors were introduced into the confirmatory factor analysis. One second-order
factor represented Perceived Behavioral Coping Ability and was reflected in
loadings allowed on social impact efficacy, general assertiveness, social
resources, dating competence, and leadership style. The other second-order
factor reflected Perceived Cognitive Control Coping Ability with hypothesized
factor indicators of social impact efficacy (the only first-order construct to load
on both second-order factors), a lack of depression, purpose in life. and a lack of
perceived loss of control.
A model was tested which included the two second-order factors as defined
above. Perceived Behavioral Coping Ability and Perceived Cognitive Control
Coping Ability were allowed to correlate freely. Three additional empirically-
determined correlations were included between pairs of first-order factor resid-
uals: Depression and leadership style, social impact efficacy and leadership
style, and depression and perceived loss of control. The factor residual of
purpose in life was fixed at zero in order to prevent it from being estimated as
negative.
This model adequately fit the data and was not significantly different from
the first-order confirmatory factor models, even though 16 fewer parameters
were necessary to represent the first-order latent factor intercorrelations (see
summary of fit indices in Table 4). This model is graphically depicted in
52 w. 1. McCarthy & M. D. Newcomb
Figure 2 omitting the observed variables for clarity. Parameter estimates are
standardized and residual variables are variances.
Table 4
Summary of Fit Statistics
Model chi2 Degrees of p Value Normed

Freedom Fit Index
Derivation Sample (N = 370)
1. Initial CFAa 480.17 224 <.001 .86

2. Final CFAb 191. 97 196 .76 .95
Model 1-2 difference 288.20 28 <.001
3. Two second-order factors
on Model 2 211.13 212 .50 .94
Model 3-2 difference 19.16 16 .26
4. Two second-order factors
on Modell 501.04 240 <.001 .87
Model 4-1 difference 20.87 16 .18
5. Second-order factors
correlated at unity 278.89 213 .001 .92
Model 5-3 difference 67.76 1 <.001
Cross-Validation Samples
6. Separate cross-validation
sample-Model 3 358.54 212 <.001 .89
7. Males only,
Model 3 219.29 212 .35 .90
8. Females only,
Model 3 352.51 212 <.001 .92
Note. a CFA = Confirmatory factor analysis. b Modified by adding 22 correlated

residuals and 5 nonhypothesized factor loadings.
Perceived Behavioral Coping Ability and Perceived Cognitive Control

Coping Ability were correlated .66, indicating a moderate association between
them (44% common variance), while retaining their own uniqueness. This cor-
relation resembled the correlation of .44 that Folkman and Lazarus (1980)
observed between emotion-focused and problem-focused coping and the range of
correlations (.48-.37) between perceived coping and cognitive control efficacy
that Ozer and Bandura (1990) recently reported. (These studies measured var-
iables rather than latent constructs, which might account for the smaller size).
The largest additional correlation was between the residuals of Social Impact
Efficacy and Leadership Style, indicating that the two second-order factors did
not account for the entire association between these two constructs. This
association may reflect an additional second-order factor for these two constructs,
over-and-above the relationship accounted for by the Perceived Behavioral
Coping Ability factor. A similar possibility may exist for Depression and
Perceived Loss of Control. These were not tested because two-indicator factors
tend to be very unstable, and the fit and interpretability of the model seem to be
quite good as it stands.
In order to determine whether the two second-order factors of self-efficacy
that we have identified are in fact separate constructs, an additional more restrict-
ed model was tested. In this model the correlation between the two second-order
factors was fixed at 1.0, operationalizing the hypothesis that they are assessing
the same quality. This model did not accurately reflect the data and was signifi-
cantly worse when compared to the previous model which allowed the two
second-order constructs to be unique (see summary of fit statistics in Table 4).
Finally, we tested the second-order factor model in the initial confirmatory
model that did not include the additional five factor loadings nor the 22 corre-
lated uniquenesses. All significant relationships were retained and the resultant
model was not significantly different from the initial model. Thus, we conclude
that the model depicted in Figure 2 is an accurate portrayal of the data that is not
biased or distorted due to model modifications.
Cross-Validation of the Second-Order Factor Model

This final model was tested in the separate, untouched random half of the
total sample, as well as the separate samples of men and women. Fit indices for
these runs are given in Table 4 (Models 6, 7, and 8). In each of these three sam-
ples, all hypothesized factor loadings and second-order factor results were
significant.
Although only the male sample fit the model according to the p-value crite-
rion, all three models had NFIs greater than .89, indicating that each fit the data
reasonably well. Similarly. the ratio of chi 2 to degrees of freedom was
consistently under 2.0, also reflecting an excellent degree of fit. All hypothesized
factor loadings on the first-order factors were significant in each of the three
sample partitions. Table 5 presents the standardized parameter estimates for the
second-order factors for the derivation sample, as well as the cross-validation,
female, and male samples. Although the magnitude of the factor loadings varied
somewhat, the general patterns of association were remarkably similar. The cor-
relation between the two second-order factors was consistently in the .64 to .66
range. Using the cross-validation method suggested by Cudeck and Browne
(1983), the specific model developed in the derivation sample was used in the
cross-validation sample (by imposing identical parameterization and parameter
estimates) and accounted for 86% of the variance of their new data. Although the
fit was worse in this second sample, the decrement was not substantial, and more
54 w. 1. McCarthy & M. D. Newcomb
important, the substantive conclusions (i.e., interpretation of parameter estimates)

was virtually identical in the derivation and cross-validation sample, when
parameters were estimated freely (Table 5). As a result, we conclude that the
factor model presented in Figure 2 is equally representative of men and women,
and thus does not differ by sex of the subject, and has been cross-validated in a
separate sample.
Table 5
Summary of Second-Order Factor Parameters for Several Sample Partitions
Parameter Derivation Cross-Validation Female Male

Sample Sample Sample Sample
Second-Order Factor Loadings
Behavioral Coping Efficacy

Social impact efficacy AO .28 .32 .31
General assertiveness .73 .74 .74 .66
Social resources .57 A9 .60 A7
Dating competence .81 .64 .69 .81
Leadership style .50 .50 .56 A6
Cognitive/Emotional Coping
Efficacy
Social impact efficacy .51 .59 .55 .59
Depression -.53 -A7 -A7 -.56
Purpose in life 1.00 .97 1.00 1.00
Perceived loss of control -.85 -.93 -.91 -.85
Factor Correlations
Perceived Behavioral Coping .66 .64 .66 .66

Ability with Perceived
CognitivelEmotional Ability
Depression (R)a with perceived
loss of control (R) AI .25 .35 .32
Social impact efficacy, (R)
with leadership style (R) A8 A9 .52 .31
Depression (R) with
leadership style (R) -.20 -.12 -.15 -.17
Note. a (R) denotes factor residual.

Cognitive/BehavioraJ Efficacy 55
Social
.30 Impact
Self-
Efficacy
.,40
Behavioral
General .73
.47 Coping
Assertiveness
Ability
.57.
Social
.67
Resources
.66
.72 Depression
(CES-D) -.53
.51
Cognitive
.48 .20 .41 Purpose 1.00
.28 Control
in Life
Ability
.28
Perceived
.28 Loss of
Control
.81
Dating
.35 Competence
.50
Leadership
.73
Style
Figure 2 Final second-order factor model for the derivation sample. The large circles are
latent constructs (the two on the right-hand side are second-order factors); the
small circles represent factor residuals. Two-headed arrows are correlations.
Parameter estimates are standardized and residual variables are variances.
Significance levels were determined by critical ratios (**p < .0 I; ***p < .00 I).
DISCUSSION
Affirmation of the Distinction Between Perceived Ability to

Self-Regulate Cognitively and Perceived Ability to Cope Behaviorally
With Environmental Challenges
Our results confirm the validity of distinguishing between beliefs about one's
ability to regulate cognitions in response to challenges associated with accomp-
lishing desired goals and beliefs about one's ability to have an impact on the
environment to accomplish desired goals. Those constructs that are primarily
intrapsychic, such as depression, purpose in life, and perceived loss of control
loaded heavily on the Perceived Cognitive Control Ability factor, but did not load
on the Perceived Behavioral Coping Ability factor. Those constructs that con-
cern active involvement with one's social environment, such as general asser-
tiveness, leadership, and dating competence, on the other hand, loaded heavily on
the Perceived Behavioral Coping Ability factor, but not on the Perceived
Cognitive Control factor.
The empirical results obtained in this study confirm the usefulness of a theo-
retical distinction that is already current and frequently applied in the literature
on coping with stress. This report, of course, goes beyond the literature on coping
to justify elaborating the self-efficacy construct to include the distinction between
perceived cognitive control and perceived behavioral abilities. The findings from
a recent prospective study involving these two distinguishable percepts of ability
(Ozer & Bandura, 1990) further affirm the importance of distinguishing between
cognitive control and behavioral coping ability. Ozer and Bandura noted in this
study of women mastering self-defense skills that perceived [behavioral) coping
self-efficacy developed more rapidly during the period of training than did cogni-
tive control self-efficacy. One explanation for this finding is that the behavioral
and attitudinal changes engendered by 22 contact hours of training in thwarting
simulated physical assaults were too many, too disparate, and too emotionally
disturbing to be fully apprehended and integrated at the time of skill acquisition.
The resulting dissociation of perceived cognitive control and perceived
[behavioral) coping self-efficacy during the skill acquisition phase was tempo-
rary. By the 6-month follow-up the positive correlation between cognitive
control efficacy and [behavioral) coping self-efficacy was restored (r = .37),
despite sustained increases in self-defense efficacy and sustained decreases in
perceived vulnerability and anxiety. It is as if time and effort were required to
alter existing patterns of cognitive control following sudden major changes in
perceived behavioral coping self-efficacy. If these findings are generalizable to
other efforts at mastering complex behaviors, they suggest the need to examine in
more detail the nature and sequencing of self-efficacy beliefs prior to, during, and
following the acquisition of mastery.
Health Consequences of High Intrapsychic Efficacy

Relative to Problem-Focused Efficacy
Individuals high in Perceived Cognitive Control Coping Ability, relative to
Perceived Behavioral Coping Ability, may end up merely palliating problems
that could be resolved behaviorally. For example, a woman suffering from an
unhappy marriage may successfully reduce that unhappiness via psychotherapy
that boosts her self-esteem, thereby increasing her ability to avoid intrusive, self-
denigrating thoughts, without doing anything about increasing her perceived
ability to influence her husband's insensitivity to her needs.
Individuals who perceive that they have little control over their environment
can nevertheless increase their probability of achieving a healthful lifestyle
change by maximizing their Perceived Cognitive Control. An example would be
an individual working in a tobacco processing plant who would like to quit
smoking. Such an individual would find few opportunities for changing the
social and physical environment to make it more conducive to quitting smoking,
but might nevertheless embark on a plan to become an exsmoker by relying on
primarily cognitive control strategies for coping with temptations to smoke.
Cognitive strategies such as thought management, self-reward, and distractive
thoughts for reducing urges to smoke could be employed to accomplish a suc-
cessful transition to being a nonsmoker even though the individual's Perceived
Behavioral Coping Ability was low.
Conversely, smokers who feel that they have little "willpower" and who are
hostile to conventional intrapsychic approaches to behavior change can never-
theless be successfully counseled to effect a successful change to being a non-
smoker by being encouraged to adopt behaviors seriatim that cumulatively are
antithetical to the smoking habit. A program of steadily increasing physical
activity, for instance, is likely to improve one's chances of becoming a long-term
nonsmoker because aerobic exercise is inherently inconsistent or incompatible
with the cigarette smoking lifestyle (e.g., Koplan, Powell, Sikes, Shirley, &
Campbell, 1982). Successful adoption of behaviors that are inconsistent with
smoking should lead to increased access to images of the self as a nonsmoker,
with an attendant increase in perceived ability to abstain from smoking (Kazdin,
1979).
Implications of the Distinction/or Interventions Designed to Change
Perceptions 0/ Self-Efficacy
Interventions designed to influence individuals' self-efficacy with respect to
a desirable health-related lifestyle change have not distinguished between
Perceived Behavioral Coping Ability and Perceived Cognitive Control Ability.
In the few instances where distinctions have been made between the use of cogni-
tive or behavioral coping strategies, there have been notable differences in the
contribution of behavioral and cognitive coping to explaining the behavior
change. The literature on the behavior change involved in drug use cessation, to
take one example, includes reports of contrasts between behavioral and cognitive
methods of coping with temptations to return to drug use. One such applied
study found that recidivists among would-be exsmokers reported relying more
heavily on behavioral coping than did the more successful exsmokers who con-
tinued to abstain from smoking (Shiffman, Reed, Maltese, Rapkin, & Jarvik,
1985). Another study investigated the determinants of cessation of heroin use
and found a similar advantage for cognitive coping relative to behavioral coping
strategies (Chaney & Roszell, 1985).
The findings reported above could be an artifact of the intervention model
used, namely the Relapse Prevention model (Marlatt & Gordon, 1985). Much of
the work on relapse prevention has focused on how to equip individuals with
self-efficacy percepts that would help them cope in situations that pose a high
risk of recidivism. The focus however, has not been on increasing the individ-
ual's perceived ability to reduce exposure to high-risk situations, but rather on
the individual's perceived ability to reduce the experience of stress in high-risk
situations. In other words, the focus has been such as to exaggerate the
importance of perceived cognitive control for coping with high-risk situations
relative to the importance of perceived ability to respond behaviorally for
avoiding or escaping from high-risk situations.
The superiority of enhancing percepts of Behavioral Coping Ability rather
than enhancing percepts of Cognitive Control Coping Ability in lifestyle change
programs is suggested by some multi-year follow-ups of heroin addicts. In their
12-year follow-up of the effects of treatment of 405 black and white male opiate
addicts, Simpson, Joe, Lehman, and Sells (1986) concluded that the most predic-
tive determinants of long-term continued abstinence were primarily behavioral:
Avoiding old drug-using friends and old hangouts, developing new friendships
with nonusers, and establishing new family ties and new work habits. In their
review of the determinants of spontaneous remission from substance use, Stall
and Biernacki (1986) arrived at similar conclusions. These results suggest that
would-be ex-addicts with strong beliefs about their ability to cope behaviorally
will experience higher rates of long-term abstinence than would-be ex-addicts
who may have strong beliefs in their ability to cope cognitively but weak beliefs
about their ability to cope behaviorally.
Percepts of ability have been shown to be important determinants of effort
and achievement (Bandura, 1986). Failure to distinguish between perceptions of
cognitive control ability and perceptions of behavioral coping ability, however,
could mask important information about the processes by which actions, beliefs
and perceptions of ability influence each other. Two examples are given below
where potentially important applications of social cognitive theory may be limit-
ed by the failure to make this distinction.
Conjecture Relating to Adolescent Maturation and Risk of Drug Abuse

An important, unexplained phenomenon in the literature on drug abuse onset
is the "window of vulnerability," namely, the relatively few teenage and young
Cognitive/BehavioraJ Efficacy 59
adult years when individuals are at risk of adopting a drug abusing lifestyle
(Abelson, Fishburne, & Cis in, 1980; Johnston, O'Malley, & Eveland, 1978;
Kandel & Logan, 1984). Lifestyle drug abuse rarely begins earnestly before
adolescence and almost never manifests de novo after age 25. The "Just Say No"
drug prevention program is premised on the belief that success in dissuading
teenagers from starting drug abuse during the teenage years will prevent drug
abuse at any age.
Why, over a lifespan of 72-78 years, should the average American only be at
risk of lifestyle drug abuse between the ages of 13 and 25, with peak onset during
the high school years? Part of the answer may be facilitated by distinguishing
between cognitive control and behavioral coping ability. The behavioral coping
ability of children is generally limited to secondary control (e.g., Rothbaum &
Weisz, 1989) because of their societally-mandated dependency on their parents
and because of their lack of life skills. Children's maturation is marked more by
increases in their cognitive control ability (e.g., distractive thoughts) than in their
behavioral coping skills (e.g., progressive goal-setting to achieve mastery over
challenge; Altshuler & Ruble, 1989).
The transition from childhood to adulthood is almost inevitably accompanied
by increases in behavioral coping ability. The life skills that are acquired include
decision-making skills, communication skills, dating skills, and employment
skills. At the beginning of the transition, these behavioral skills are uniformly
absent but young adolescents become increasingly aware of the need to acquire
them (Katz & Zigler, 1967). There is considerable distress and anxiety that
accompanies adolescents' increasing realization of the need for life skills in the
immediate absence of their acquisition. This distress and anxiety are palliated in
teenagers performing well in school by societally-administered reassurances that
their career trajectory is favorable and that, by implication, the teenagers need not
fear a characterological inability to acquire the necessary life skills. For these
success-bound teenagers, positive self-statements are easily accessible as anti-
dotes to the inevitable anxiety that their immaturity occasions. For many
teenagers not performing well in school and otherwise not receiving societal reas-
surances concerning future expectations of success, however, only the actual
acquisition of life skills will permanently reduce the fear that they will never be
fully accepted as autonomous, responsible adults. The literature, shows, in fact,
that drug abuse-prone teenagers are characterized by a syndrome of "accelerated
maturity," (Gritz, 1977), which manifests in precocious sexual behavior, mar-
riage, cessation of schooling, and employment. Despite the uniform absence of
life skills at the beginning of adolescence, only a minority go on to adopt a life-
style habit of drug abuse. The at-risk teenagers who successfully avoid drug
abuse are those who can through cognitive control alone reduce their immaturity-
associated anxiety to acceptable levels. At-risk teenagers who successfully avoid
drug abuse tend to come from intact families, suggesting that family social sup-
port can strengthen self-percepts of ability to control immaturity-associated anxi-
ety. For at-risk teenagers without the requisite cognitive control skills, their
immaturity-associated anxiety is functionally (but only intennittently) palliated

by regular administration of psychoactive drugs, especially nicotine and alcohol.
Kaplan, Martin, and Robbins (1982) demonstrated prospectively that non-drug
abusing teenagers with low self-esteem were significantly more likely in future
years to become drug abusers than their non-drug abusing agemates with high
self-esteem. Orug abuse, therefore, can be viewed as a functional way to medi-
cate for intrapsychic discomfort. By the time of onset of adulthood, most
individuals have demonstrated successful mastery of at least the rudiments of the
most important life skills. With the ebbing of immaturity-associated anxiety,
there is decreased need for psychoactive agents to provide functional relief. By
the time of young adulthood, unfortunately, many individuals have become phy-
siologically dependent on their chosen drugs and cannot, therefore, stop using the
drug just because the original need for the drug has disappeared.
A Conjecture Concerning Self-Efficacy Gender Differences
in Sex-Role Socialization
The distinction between perceived cognitive control ability and perceived
behavioral coping ability may similarly shed light on the origins of observed dif-
ferences between men and women in mastery of a variety of life's challenges,
including occupational achievement (Austin & Hanisch, 1990) and weight con-
trol (Jeffery, French, & Schmid, 1990). Women's continuing preference for
teaching, nursing, and childcare, and men's continuing preference for construc-
tion, community safety (policy, fire, paramedics), and surgery, are consistent
with women relying more on cognitive control coping and men relying more on
behavioral coping. Similarly, women's relative reluctance to adopt increased
physical activity and their preference for relying on willpower relative to men as
a strategy for maintaining desirable weight is also consistent with women relying
on cognitive control and men relying on behavioral coping. This difference
between men and women may help to explain why women to view weight as less
controllable than men (Jeffery et aI., 1990), given the clear long-tenn advantage
that exercise represents as a weight loss strategy (King, Frey-Hewitt, Oreon, &
Wood, 1989; Koplan et aI., 1982).
In her review of the literature on sex role socialization, Weitz (1977) cited
studies of adult communication patterns in same-sex groups in which it was
observed that women tended to be socioemotional whereas men tended to be
task-oriented. Although similar communication patterns were not observed in
children, Weitz noted a consistent association of activity and aggression in boys
and not in girls. Her evidence suggested that boys were encouraged to combat
their frustrations behaviorally and that girls were more encouraged to palliate
their frustrations through cognitive control strategies. Consistent gender dif-
ferences in adult perfonnance could well have roots in the different coping
strategies that boys and girls are encouraged to develop.
CONCLUSION
Further research on how perceived cognitive control ability and perceived
behavioral coping ability vary among individuals, among situations, and within
individuals over time seems warranted. Investigating the relative importance of
perceived cognitive control and perceived behavioral coping ability in therapy-
mediated lifestyle change and spontaneous, unaided lifestyle change would seem
especially worthwhile. This distinction would also seem useful in illuminating
more clearly why there exist differences in mastery between men and women,
and in better understanding what contributes to the youthful decision to adopt a
drug abusing lifestyle.
REFERENCES
Abelson, H., Fishburne, P., & Cisin, I. (1980). The national survey on drug abuse:
Main findings 1979. Rockville, MD: National Institute on Drug Abuse.
Altshuler, 1. L., & Ruble, D. N. (1989). Developmental changes in children's
awareness of strategies for coping with uncontrollable stress. Child Development, 60,
1337-1349.
Austin, 1. T., & Hanisch, K. A. (1990). Occupational attainment as a function of abili-
ties and interets: A longitudinal analysis using project TALENT data. Journal of
Applied Psychology, 75,77-86.
Bachman, 1. G., O'Malley, P. M., & Johnston, L. D. (1984). Drug use among young
adults: The impacts of role status and social environment. Journal of Personality
Bandura. A. (1977). Self-efficacy: Toward a unifying theory of behavioral change.
Psychological Review, 84,191-215.
Bandura, A. (1986). Social foundations of thought and action: A social cognitive
theory. Englewood Cliffs, NJ: Prentice-Hall.
Bandura, A., & Wood, R. (1989). Effect of perceived controllability and performance
standards on self-regulation of complex decision making. Journal of Personality and
Social Psychology, 56, 805-814.
Bentler, P. M. (1980). Multivariate analysis with latent variables: Causal modeling.
Annual Review of Psychology, 31, 419-456.
Bentler, P. M. (1983). Some contributions to efficient statistics in structural models:
Specification and estimation of moment structures. Psychometrika, 48, 493-517.
Bentler, P. M. (1986). Structural modeling and Psychometrika: An historical per-
spective on growth and achievements. Psychometrika, 51, 35-51.
Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness of fit in the
analysis of covariance structures. Psychological Bulletin, 88, 588-606.
Bentler, P. M., & Chou, C. P. (1986, April). Statistics for parameter expansion and
construction in structural models. Paper presented at the American Educational Re-
search Association meeting, San Francisco.
Bentler, P. M., & Newcomb, M. D. (1978). Longitudinal study of marital success and
failure. Journal of Consulting and Clinical Psychology, 46, 1053-1070.
Bentler, P. M., & Newcomb, M. D. (1986). Personality, sexual behavior, and drug use
revealed through latent variable methods. Clinical Psychology Review, 6, 363-385.
Blatt, S. 1., Quinlan, D. M., Chevron, E. S., McDonald, c., & Zuroff, D. (1982). De-
pendency and self-criticism: Psychological dimensions of depression. Journal of
Consulting and Clinical Psychology, 50, 113-124.
Chaney, E. F., & Roszell, D. K. (1985). Coping in opiate addicts maintained on metha-
done. In S. Shiffman & T. A. Wills (Eds.), Coping and substance use (pp. 267-291).
Orlando, FL: Academic Press.
Clark, V. A., Aneshensel, C. S., Frerichs, R. R., & Morgan, T. M. (1981). Analysis of
effects of sex and age in response to items on the CES-D scale. Psychiatry Research,
5,171-181.
Courtois, C. A. (1988). Healing the incest wound. New York: Norton.
Crumbaugh, 1. C. (1968). Cross-validation of purpose in life test based on Frankl's
concepts. Journal of Individual Psychology, 24, 74-81.
Crumbaugh, 1. C., & Maholick, L. T. (1964). An experimental study in existentialism:
The psychometric approach to Frankl's concept of neogenic neurosis. Journal of
Crumbaugh, J. C., & Maholick, L. T. (1969). Manual of instructions for the purpose in
life test. Munster, IN: Psychometric Affiliates.
Cudeck, R., & Browne, M. W. (1983). Cross-validation of covariance structures.
Multivariate Behavioral Research, /8, 147-167.
Donovan, J. E., Jessor, R., & Jessor, L. (1983). Problem drinking in adolescence and
young adulthood: A follow-up study. Journal of Studies on Alcohol, 44,109-137.
Folkman, S., & Lazarus, R. S. (1980). An analysis of coping in a middle-aged com-
munity sample. Journal of Health and Social Behavior, 21,219-239.
Gritz, E. R. (1977). Smoking: The prevention of onset. In Jarvik, M. E. et al.
Research on smoking behavior (pp. 290-307). NIDA Research Monograph #17,
DHEW Pub. No. (ADM) 78-581.
Harlow, L. L., Newcomb, M. D., & Bentler, P. M. (1987). Purpose in life test as-
sessment using latent variable methods. British Journal of Clinical Psychology, 26,
235-236.
Huba, G. J.. & Bentler, P. M. (1982). A developmental theory of drug use: Derivation
and assessment of a causal modeling approach. In B. P. Baltes, & O. G. Brim, Jr.
(Eds.), Life-span development and behavior, Volume 4 (pp. 147-203). New York:
Academic Press.
Huba, G. J., & Bentler, P. M. (1983). Causal models of the development of law
abidance and its relationship to psychosocial factors and drug use. In W. S. Laufer &
J. M. Day (Eds.), Personality theory, moral development, and criminal behavior (pp.
165-215). Lexington, MA: Heath.
Husaini, B. A., Neff, J. A., Harrington, J. B., Hughes, M. D., & Stone, R. H. (1980).
Depression in rural communities: Validating the CES-D scale. JournalofCommun-
ity Psychology, 7, 137-146.
Jeffery, R. W., French, S. A., & Schmid, T. 0. (1990). Attributions for dietary failures:
Problems reported by participants in the Hypertension Prevention Trial. Health Psy-
chology, 9, 315-329.
Johnston, L., O'Malley, P., & Eveland, L. (1978). Drugs and delinquency: A search
for causal connections. In D. B. Kandel (Ed.), Longitudinal research on drug use:
Empirical findings and methodological issues (pp. 137-156). Washington, DC:
Hemisphere-Wiley.
Kandel, D. B. (1984). Marijuana users in young adulthood. Archives of General
Psychiatry, 4/, 200-209.
Kandel, D. B., & Logan, J. A. (1984). Patterns of drug use from adolescence to young
adulthood: I. Periods of risk for initiation, continued use, and discontinuation.
American Journal of Public Health, 74, 660-666.
Kaplan, H. B., Martin, S. S., & Robbins, C. (1982). Applications of a general theory of
deviant behavior: Self-derogation and adolescent drug use. Journal of Health and
Social Behavior, 23,274-294.
Katz, P., & Zigler, E. (1967). Self-image disparity: A developmental approach.
Kazdin, A. E. (1979). Imagery elaboration and self-efficacy in the covert modeling
treatment of unassertive behavior. Journal of Consulting and Clinical Psychology,
47, 725-733.
King, A. C, Frey-Hewitt, B., Dreon, D. M., & Wood, P. D. (1989). Diet vs. exercise in
weight maintenance. The effects of minimal intervention strategies on long-term
outcomes in men. Archives of Internal Medicine, 149,2741-2746.
Koplan, J. P., Powell, K. E., Sikes, R. K., Shirley, R. W., Campbell, C. C (1982). An
epidemiologic study of the benefits and risks of running. Journal of the American
Medical Association, 248, 3118-3121.
Levenson, R. W., & Gottman, 1. M. (1978). Toward the assessment of social compe-
tence. Journal of Consulting and Clinical Psychology, 46,453-462.
Marlatt, G. A., & Gordon, J. (1985). Relapse prevention: Maintenance strategies in
addictive behavior change. New York: Guilford.
Miller, J. D., Cisin, I. H., Gardner-Keaton, H., Harrell, A. V., Wirtz, P. W., Abelson,
H. I., & Fishburne, P. M. (1983). National survey on drug abuse: Mainfindings 1982.
PHS, NIDA, DHHS Pub. No. (ADM) 83-1263. Washington, DC: U. S. Government
Printing Office.
Newcomb, M. D. (1984). Sexual behavior, responsiveness, and attitudes among
women: A test of two theories. Journal of Sex & Marital Therapy, 10,272-286.
Newcomb, M. D. (1986). Nuclear attitudes and reactions: Associations with depres-
sion, drug use, and quality of life. Journal of Personality and Social Psychology, 50,
906-920.
Newcomb, M. D., & Bentler, P. M. (1986). Loneliness and social support: A confirma-
tory hierarchical analysis. Personality and Social Psychology Bulletin, 12,520-535.
Newcomb, M. D., & Bentler, P. M. (1987). Self-report methods of assessing health
status and health service utilization: A hierarchical confirmatory analysis. Multi-
variate Behavioral Research, 22, 415-436.
Newcomb, M. D., & Bentler, P. M. (1988a). Consequences of adolescent drug use:
Impact on the lives of young adults. Newbury Park, CA: Sage.
Newcomb, M. D., & Bentler, P. M. (l988b). Impact of adolescent drug use and social
support on problems of young adults: A longitudinal study. Journal of Abnormal
Newcomb, M. D., & Harlow, L. L. (1986). Life events and substance use among adole-
scents: Mediating effects of perceived loss of control and meaninglessness in life.
O'Leary, A. (1985). Self-efficacy and health. Behavior Research and Therapy, 23,
437-451.
Ozer, E. M., & Bandura, A. (1990). Mechanisms governing empowerment effects: A
self-efficacy analysis. Journal of Personality and Social Psychology, 58, 472-486.
Pearlin, L., & Schooler, C (1978). The structure of coping. Journal of Health and
Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in
the general population. Applied Psychological Measurement, I, 385-401.
Roberts, R. E. (1980). Reliability of the CES-D scale in different ethnic contexts.
Psychiatry Research, 2, 125-134.
Rothbaum, F., & Weisz, J. R. (1989). Child psychopathology and the quest for control.
Newbury Park, CA: Sage.
Ryckman, R. M., Robbins, M. A., Thornton, B., & Cantrell, P. (1982). Development
and validation of a physical self-efficacy scale. Journal of Personality and Social
Shiffman, S., Read, L., Maltese, J., Rapkin, D., & Jarvik, M. E. (1985). Preventing
relapse in ex-smokers: A self-management approach. In A. Marlatt & 1. Gordon
(Eds.), Relapse prevention (Chapter 8, pp. 472-520). New York: The Guilford Press.
Simpson, D. D., Joe, G. W., Lehman, W. E., & Sells, S. B. (1986). Addiction careers:
Etiology, treatment, and 12-year follow-up outcomes. Journal of Drug Issues, /6,
107-121.
Stall, R., & Biernacki, P. (1986). Spontaneous remission from the problematic use of
substances: An inductive model derived from a comparative analysis of the alcohol,
opiate, tobacco, and food/obesity literatures. International Journal of the Addictions,
21,1-23.
Stein, 1. A., Newcomb, M. D., & Bentler, P. M. (1986). Stability and change in per~0n
ality: A longitudinal study from early adolescence to young adulthood. Journ.1! u}
Strecher, V. J., DeVellis, B. M., Becker, M. H., & Rosenstock, 1. M. (1986). The 10k
of self-efficacy in achieving health behavior change. Health Education Quartel/y,
/3,73-91.
Taylor, S. E. (1986). Health psychology. New York: Random House.
Weissman, M. M., Shokomskas, D., Pottenger, M., Prusoff, B. A., & Locke. B. Z.
(1977). Assessing depressive symptoms in five psychiatric populations: A validation
study. American Journal of Epidemiology, /06, 203-214.
Weitz, S. (1977). Sex roles. New York: Oxford University Press.
Wood, R., & Bandura, A. (1989). Impact of conceptions of ability on self-regulatory
mechanisms and complex decision making. Journal of Personality and Social
Author Notes
This research was supported by grant DA01070 from the National Institute
on Drug Abuse. The computer assistance of Sandy Yu and production assistance
of Julie Speckart are gratefully acknowledged.
EXPECTANCIES AS MEDIATORS
BETWEEN RECIPIENT
CHARACTERISTICS AND
SOCIAL SUPPORT INTENTIONS
Ralf Schwarzer, Christine Dunkel-Schetter,

Bernard Weiner, and Grace Woo
It is assumed that the motivation to extend social support is governed

by specific emotions and cognitions, among them outcome expectan-
cies and self-efficacy expectancies. Two experiments were conducted
to explore this assumption, Study I dealing with outcome expectancy
and Study /I dealing with self-efficacy expectancy. In Study I, out-
come expectancies toward eight disease-related stigmas and the in-
tention to extend social support were examined with two experimental
conditions. The onset of the stigmas was varied as being either con-
trollable or uncontrollable. In addition, the target person was describ-
ed either as actively coping with the stigma or as not actively coping.
Examined were the effects oj onset controllability and coping on pity,
outcome expectancy, and willingness to support the target person. In a
within-groups design, 84 subjects were confronted with all eight stig-
mas under four different conditions. Both experimental factors influ-
enced the reported reactions. The coping variable appeared to be
stronger than the controllability variable and. in addition, outcome
expectancy was a somewhat more important mediator of helping than
pity. However, the pattern of data was context-specific, i.e., different
sets of predictors emerged for different stigmas. Study /I was a simi-
lar experiment pursuing the notion that the motivation to help is af-
fected by the belief that one can be effective as a helper (self-efficacy
expectancy). It examined whether self-efficacy expectancy for helping
a rape victim served as a mediator of the relationship between recipi-
ent characteristics and support intentions. The recipient characteris-
tics assessed were victim coping and controllability of the assault.
Both pity and self-efficacy expectancy emerged as good predictors of
support, whereas controllability and coping were of lesser influence.
According to Bandura's cognitive-social theory, human behaviors are partly

governed by expectancies, in particular by outcome expectancies and self-
efficacy expectancies (Bandura, 1977, 1986, 1991). Many studies, some of them
66 R. Schwarzer et al.
presented in this volume, have applied this assumption to specific behaviors in

various domains of human functioning such as achievement, organizational man-
agement, or health. There seems to be, however, no application to studies on
social support. The willingness to help others depends partly on one's emotions
at the time, but helping also depends on judgments about the specific situation,
characteristics of the recipient, and one's self. Among such cognitions are expect-
ancies about the likelihood that the situation can be changed and regarding one's
ability to provide the necessary social support. Expecting a condition to improve
under certain circumstances represents an outcome expectancy. Belief in oneself
as an effective support provider in a particular situation represents a self-efficacy
expectancy. These cognitions are hypothesized to serve as causal mediators of
the relationships between antecedent recipient characteristics and consequent
intentions to extend social support. In addition, a number of other factors outlined
below are considered important in the study of social support.
The present chapter reports two studies. The first one deals with the mediat-
ing role of outcome expectancy, the second one with the mediating role of self-
efficacy expectancy. In the following sections, we describe in more detail the
constructs involved in this research, in particular perceived controllability,
perceived coping, expectancies, and social support.
Perceived Controllability
Attribution theory has recently been extended to the study of social stigmas
and reactions to the stigmatized (Weiner, Perry, & Magnusson, 1988). By social
stigma we mean a discrediting condition or mark that defines a person as
"deviant, flawed, limited, spoiled, or generally undesirable" (Jones et aI., 1984,
p. 6). Among others, physical deformities, behavioral problems such as excessive
eating and drinking, and diseases can be regarded as stigmas. Attribution theory
is relevant to the study of stigmas because individuals typically search for the
cause(s) of a negative state or condition existing in others. That is, observers con-
fronted with a "markable" target initiate an attributional search to determine the
origin of the stigma.
Researchers have identified controllability as one of the basic dimensions of
perceived causality (Weiner, 1985, 1986). Controllable causes are those which an
actor can volitionally change, whereas uncontrollable causes are not subject to
personal mastery or management. The onset of a drug problem, for example, is
seen as controllable if a person has been experimenting with drugs out of curi-
osity, whereas it is perceived as comparatively uncontrollable if a person has had
medical treatment with drugs and thereby developed a dependency (Weiner et aI.,
1988). In a similar manner, the onset of a heart disease is construed as control-
lable if the person has led an unhealthy life-style, including smoking and a poor
diet, whereas it is considered relatively uncontrollable if hereditary factors have
played a major role in the illness.
Expectancies and Help Intentions 67
Affective and Behavioral Reactions

It has been documented that the perceived controllability of a social stigma
determines disparate affective reactions toward the target person and different
behavioral responses as well (e.g., Dejong, 1980; Weiner et al., 1988). More
specifically, uncontrollable origins of stigmas tend to elicit pity and offers of
help, whereas controllable origins tend to elicit anger and no help (see
Reisenzein, 1986; Schmidt & Weiner, 1988; Weiner et al., 1988). Hence, it has
been shown that experimentation with drugs and an unhealthy life-style as causes
of stigmas yield much anger, little pity, and relative neglect, whereas drug prob-
lems due to medical treatment, and heart disease derived from genetic factors,
give rise to little anger, much pity, and prosocial responses.
Perhaps more than in any other area within the field of social motivation,
investigators of helping behavior have assumed that emotions play an important
motivational role (see review in Carlson & Miller, 1987). These emotions have
included discomfort (e.g., Cialdini, Darby, & Vincent, 1973), distress (e.g.,
Batson, O'Quinn, Fultz, Vanderplas, & Isen, 1983), empathy (e.g., Batson, 1990;
Hoffman, 1975), gratitude (e.g., Goranson & Berkowitz, 1966), guilt (e.g.,
Hoffman, 1982), as well as pity and anger (Schwarzer & Weiner, 1990, 1991).
While there is strong support for an attribution-emotion-helping link, there are
also studies that have failed to demonstrate this effect. Capitalizing on a real-life
event, Amato, Ho, and Partidge (1984) sent survey questionnaires to residents
living near the setting of a major bushfire which killed 46 people and destroyed
over 2,000 homes. The questionnaire addressed perceptions of causality and
responsibility, affective reactions, and helping behavior. Most people reported
donating to the victims regardless of the amount of responsibility attributed to
them. The obvious high degree of need in this context seemed to have over-
powered the attribution of control effects.
Jung (1988) presented subjects with vignettes depicting a close friend experi-
encing a variety of common problems, with manipulations of the responsibility
for the problem. For each vignette, subjects rated the target person's
deservedness of fate, perception of how helpful social support would be for the
problem, and their likelihood of providing social support. Perceived deservedness
of fate was greater for those viewed as having high responsibility. Perceived
benefits of social support were also higher in this case. However, neither factor
affected the likelihood of social support provision.
Skokan (1990) examined the affective responses and support behaviors
extended towards a roommate who is dealing with either cancer or the death of
her father. Subjects were presented with scenarios which depicted the roommate
as either responsible or as not responsible for the onset of the critical event. In
her initial within-subjects analysis, controllability was associated with more
anger, less sympathy and less social support; however, when reanalyzed as a
between-subjects design because of order effects, the impact of controllability on
sympathy and support disappeared.
Perceived Coping
It remains unclear whether stigma onset, which is a distant event, is the sole
or main detenninant of affective and behavioral reactions toward the stigmatized
or whether subsequent events, controllable or uncontrollable, alter the causal
sequence. Drug experimentation and poor life-style, for example, might be weak
predictors of the emotions and behaviors of observers when compared with the
present efforts of the target person to cope with the consequences of the stigma.
In the achievement domain, it is obvious that even after failure due to lack of
effort, present expenditure of effort to compensate or recover generates positive
affect and rewards for the failed student (Karasawa, 1991; Weiner, 1985). When
generalized to the health domain, this finding suggests that positive coping
attempts with a serious health condition could play an important role in
detennining the affective and behavioral reactions of others.
Skokan (1990) distinguished in her scenario experiment between adaptive
coping and maladaptive coping. In the adaptive condition, the target person who
either had cancer or was bereaved, tried to stay optimistic and to look for ways to
go on with her life and to grow from the experience. In the maladaptive condi-
tion, she dwelled on the negative aspects of the situation and did not try to over-
come the crisis instrumentally. Adaptive coping of the target was related to less
anger in subjects but had mixed effects on their willingness to offer social sup-
port. In the bereavement condition, poor coping elicited less support, but in the
cancer condition, unexpectedly, poor coping elicited even more support.
Silver, Wortman, and Crofton (1990) studied subject reactions to a cancer
patient who was portrayed either as a "good coper," a "bad coper" or a "balanced
coper." In the good coping condition, the target person expressed an optimistic
view of her illness and appeared to be coping well. In the balanced coping condi-
tion, she conveyed distress about what was happening, but also indicated that she
was trying her best. In the poor coping condition, she displayed distress about
what was happening and appeared to have difficulty coping. In nine out of ten
comparisons, the responses to confederates who were portrayed as having posi-
tive or balanced coping styles were significantly more favorable than were
responses to poor copers.
In sum, both the origin of a problem and its solution are hypothesized to be
important when examining reactions of others toward the stigmatized person
(Brickman et aI., 1982). That is, the responsibility for causing a problem should
be separated from the responsibility for maintaining or not alleviating it. This
important distinction has been ignored in prior research on attributions (see also
Karasawa, 1991; Schwarzer & Weiner, 1991). The present studies compare the
effects of perceived onset controllability with those of perceived coping efforts
on pity, outcome expectancy, and social support towards the stigmatized and
examines the mediating role of pity and expectancy.
Expectancies
The focus of the present paper is on the role of mediating factors that link
attributions and affect regarding a social stigma to behavioral intentions or to
actual support behavior. Bandura (1977, 1986, t 991) has convincingly demon-
strated that expectancies are very important social-cognitive mediators of action.
There are two major cognitions of this kind, outcome expectancies and self-
efficacy expectancies. In the first experiment, we deal with outcome expectancies
that refer to the possibility of improvement of a condition. The subjects were
asked how likely it is that a target person's condition would improve under partic-
ular circumstances. It is hypothesized that an individual's active coping with an
ailment will trigger positive outcome expectancies in the observer. Coping
behavior implicitly refers to the stability of a stigma. If the victim is not actively
involved in alleviating the distress, maintaining functioning and moving on with
daily life, one would have little reason to expect an improvement; support may be
seen as wasted labor. If, however, a great deal of effort is expended by the victim
in solving the problem, one can expect that changes are more likely and that sup-
plementary contributions would be a worthwhile investment. This reasoning does
not apply to situations that require acceptance; that is, we are likely to help
people who behave passively when passivity is required in the situation.
In the second experiment, the focus is on self-efficacy expectancy in terms of
one's helping capabilities. Empathy, perspective taking, comforting skills and so
on, not only facilitate social support in an objective sense (Batson, 1990; Clary &
Orenstein, 1991); these abilities also have to be perceived by the help provider in
order to establish a motivation to help. Help-specific self-efficacy deals with cog-
nitions about one's capability to support others and to make a difference with this
support; it refers to one's perceived personal resources to provide competent
assistance and to achieve relief for a sufferer.
Social Support
Social support has been defined as an exchange of resources "perceived by

the provider or the recipient to be intended to enhance the well-being of the
recipient" (Shumaker & Brownell, t 984, p. t 3). This definition requires that
either the provider or the recipient must perceive that the provider has a positive
intent. Intentions have also been claimed as being the best predictors of a variety
of behaviors; this is well-documented in research based upon the Theory of
Reasoned Action (Fishbein & Ajzen, 1975) and the Theory of Planned Behavior
(Ajzen, 1988). Evidence of the influence of help intentions on actual helping
behavior has been found by Borgida, Simmons, Conner, and Lombard (1990) and
Dalbert, Montada, and Schmitt (1988). Whether intentions to help are accurately
perceived by the provider or by the recipient is a related but different question
(Dunkel-Schetter & Bennett, 1990; Dunkel-Schetter, Blasband, Feinstein, &
Bennett, 1991).
Several factors determine the likelihood that a supportive exchange actually

takes place. Stress factors, relationship factors, recipient factors, and provider
factors have been discussed and somewhat studied (Dunkel-Schetter & Skokan,
1990). We will deal here with the latter two exclusively. Recipient factors are
critical determinants of support. Victims who are not only distressed, but are also
not responsible for the event, and who invest a great deal of effort to manage
their condition, are apt to elicit more help than those who are responsible them-
selves for their misfortune and who do not take action to solve their issue
(Bennett-Herbert & Dunkel-Schetter, in press; Brickman et aI., 1982). Creating
frustration and helplessness in the potential provider leads to a lesser likelihood
of support (Dunkel-Schetter & Wortman, 1981, 1982). The expression of too
much distress strains the social network, evokes negative reactions, and turns
those away who would have been supportive if the distress level had only been
moderate. Another reason why the network may not be mobilized is if a victim is
not coping adaptively. Passive, depressive and ungrateful victims or patients are
seen as socially unattractive and, therefore, receive less support in the long run
(Barbee, 1990; Gurtman, 1986; Notarius & Herrick, 1988). Paradoxically, those
subjects who have valuable personal resources such as competence, high self-
esteem, locus of control, and optimism and who make use of their resources seem
to elicit a stronger tendency in others to extend support.
Provider factors have been intensively studied in social psychology research
on helping (Batson, 1990; Berkowitz, 1987; Dovidio, 1984; Eisenberg & Miller,
1987; Jung, 1988). It makes a difference how the cause of the problem is attrib-
uted. If it is seen as controllable then the victim is blamed and negative emotional
reactions are aroused such as anger, leading to neglect of the sufferer. If, on the
other hand, the cause is seen as uncontrollable and the person does not seem to be
responsible for the problem, then positive emotions such as pity emerge, which
make help more likely (Weiner, 1985). Thus, emotions are mediators of attribu-
tions and behavioral intentions. According to Batson (1990), empathy predicts
altruistic motivation to help, whereas a provider's distress tends to elicit egoistic
motivation, which does not induce help. These two theories by Weiner and
Batson are closely related in terms of emotional mediators of motivation. Pity can
be matched to empathy as a predictor of help, and anger parallels distress in pre-
dicting neglect. Betancourt (1990) has attempted to integrate both views by
manipulating experimentally the controllability of onset of a problem as well as
inducing different perspectives in the potential support provider. He found that
both experimental factors influenced perceived controllability and empathic
emotions that, in tum, influenced helping.
In the present chapter, the focus is on experimentally manipulated recipient
factors, but it is kept in mind that these do not operate in an isolated manner.
Rather, they interact with on-going responses by the provider during a specific
social encounter. It is only of secondary importance whether the victim is
actually responsible for the problem and whether active coping is executed.
Moreover, the degree to which the provider makes these attributions, is
considered to be critical. The perception in the beholder may be more relevant

than the actual cause of the onset of the stigma or the actual coping behavior.
The present experiments were designed to examine the effects of perceived
controllability and perceived coping on pity as an affective reaction and on
expectancies that, in turn, were hypothesized to exert an influence on support
intent. The studies differ in terms of the scenarios used and in terms of the
expectancy variables. While Study I deals with outcome expectancy, Study II
deals with self-efficacy expectancy.
STUDY I
Method
Sample. The subjects were 84 male and female students at the University of
California, Los Angeles, who received credit in an introductory psychology
course for their participation. They were randomly assigned to one of four groups
(see below) and given questionnaires in small group sessions with anonymity
assured. I
Design. Eight health-related stigmas were selected, each of which was
manipulated with respect to onset controllability and coping effort. Each subject
received four of the eight stigmas paired uniquely with one of the four control-
lability conditions (2 Levels of Onset Responsibility x 2 Levels of Coping). Sub-
jects were divided into four groups that received different combinations of
stigmas and conditions (see Table I).
Table I
Experimental Design
Onset Responsible Onsel Irresponsible
No Coping Coping No Coping Coping
Group 1 Aids Cancer Drug abuse Heart disease

Group 2 Cancer Drug abuse Heart disease Aids
Group 3 Drug abuse Heart disease Aids Cancer
Group 4 Heart disease Aids Cancer Drug abuse
Group 1 Anorexia Child abuse Depression Obesity

Group 2 Child abuse Depression Obesity Anorexia
Group 3 Depression Obesity Anorexia Child abuse
Group 4 Obesity Anorexia Child abuse Depression
1 Study I was conducted by Ralf Schwarzer and Bernard Weiner.

As shown in Table 1, one part of the design included four stigmas (AIDS,
cancer, drug abuse, and heart disease) paired with the four conditions, while a
second part replicated the flrst but used another four stigmas (anorexia, depres-
sion, obesity, and child abuse). Thus, there were two within-group factors (onset
controllability and coping) and one between-group factor (stigma set). This
design allowed for an overall analysis as well as for stigma-specific subanalyses.
Four vignettes were created for each stigma consisting of: (a) onset
responsibility and low coping; (b) onset responsibility and high coping; (c) no
onset responsibility and low coping; and (d) no onset responsibility and high
coping. As an example, the obesity vignettes are given:
1. Maladaptive coping, controllable. ¥our roommate has become excessively over-
weight, and is experiencing severe problems in social- and work-related activities.
Excessive eating and lack of exercise have been the primary contributors to the
obesity. This roommate does not take any steps to lose weight, either by dieting,
exercising or by following a medical regimen.
2. Adaptive coping, controllable. Your roommate has become excessively overweight,
and is experiencing severe problems in social- and work-related activities. Exces-
sive eating and lack of exercise have been the primary contributors to the obesity.
Recently this roommate has commenced a new diet prescribed by a physician, and
is regularly exercising.
3. Maladaptive coping, uncontrollable. ¥our roommate has become excessively over-
Glandular dysfunction has been identified as the reason for the obesity. This room-
mate does not take any steps to lose weight, either by dieting, exercising or by
following a medical regimen.
4. Adaptive coping, uncontrollable. Your roommate has become excessively over-
Glandular dysfunction has been identified as the reason for the obesity. Recently
this roommate has commenced a new diet prescribed by a physician. and is
regularly exercising.
Measures. The dependent variables were the following 9-point rating scales,
anchored with extremes such as not at all and very much so. Pity was assessed by
the single item "How much pity would you feel?"
Typically, outcome expectancies are worded in an "if-then manner." In the
present experiment, however, the if-component was given by the four experi-
mental conditions such as: "If the stigma is uncontrollable and if the victim is
actively coping with it, then ... " Because of these implicit assumptions, the mea-
surement of the outcome expectancy was restricted to the then-component and
simply worded: "How likely is it that the condition will improve?"
Social support intention was measured by seven items representing different
kinds of social support. However, this was a homogeneous scale (Cronbach's
alpha for the seven social support items was .91), and, therefore, the aggregated
score was used as an indicator of support intentions. The items were:
1. How much would you like to extend support to your roommate?

2. How much time would you be willing to spend talking and listening?
3. How much money would you be willing to donate in order to provide the
best possible treatment?
4. How much would you like to go on a holiday trip with your roommate?
5. How much would you be willing to give advice and information?
6. How much would you be willing to console and reassure your roommate
when being upset?
7. How willing would you be to assist with a small problem?
Other dependent variables were analyzed previously within the framework of
analysis of variance, and some of the results are published elsewhere. However,
we have only reported about the stigmas of heart disease (Schwarzer & Weiner,
1990), AIDS and cancer (Schwarzer & Weiner, 1991).
Results
To examine the role of pity and outcome expectancy as mediators of the rela-
tionship between victim characteristics and provider support intentions, a
structural equation model was specified with controllability and coping as exo-
genous variables and pity, expectancy, and support as endogenous variables. This
is a straightforward single indicator model with manifest variables. The two
orthogonal experimental factors were believed to influence emotions and cogni-
tions, whereas emotions and cognitions were specified to influence the behavioral
intention directly. Controllability and coping, therefore, could exert indirect
effects on support intent through pity and expectancy but were constrained not to
exert direct effects, because this would not be in line with theory or past research.
The two alternative mediating factors were pity and expectancy, and for both of
them the size of their mediating effect was computed in addition to their direct
impact on support intent (see Figures 2 to 9). This procedure was repeated eight
times, for each stigma individually. Eight path analyses were carried out with the
LISREL VII program (Joreskog & Sorbom, 1988).
First, the degree to which the experimental data fitted the structural equation
model was examined. Several indices of fit have been suggested in the literature
(cf. Bentler, 1980). We have used five of them in this study, (a) the chi-square
test which, if significant, indicates that the data deviate from the model, (b) the
chi-square Idf ratio which takes the degrees of freedom into account (df = 3) and
which should be as low as possible; ratios above 3.0 are usually seen as unsatis-
factory, (c) Joreskog's Goodness of Fit Index (GFl) which should be close to
unity, (d) his Adjusted Goodness of Fit Index (AGFl) that makes an adjustment
to the degrees of freedom and also should be as high as possible, and (e) the Root
Mean Square Residual (RMSR) which is an index derived from the deviations of
the original correlation matrix from the reproduced correlation matrix on the
basis of the estimated parameters; this index should not exceed .05.
Table 2 summarizes the results of all eight path analyses. In six of eight
cases, an excellent fit emerged, whereas the stigmas "Cancer" and "Child Abuse"
turned out to be associated with a less appropriate fit. Overall, these satisfactory
results indicate that the model specification is in line with the experimental data,
but also that the specific stigma context makes a difference.
Table 2
Goodness of Fit for the Eight Path Models
Stigma chi2 p chi 2/df OF! AOFI RMSR
AIDS 4.36 .23 1.45 .98 .90 .05

Cancer 9.49 .02 3.16 .96 .79 .07
Drug abuse 1.46 .69 0.49 .99 .97 .03
Heart disease 2.44 .49 0.81 .99 .94 .04
Anorexia 0.92 .82 0.31 .99 .98 .02
Child abuse 15.18 .002 5.06 .94 .69 .08
Depression 1.50 .68 0.50 .99 .97 .03
Obesity 4.65 .20 1.55 .98 .90 .05
Note. OFI == goodness of fit, AOFI == adjusted OFI, RMSR == root mean square residual.
Table 3
Percent of Explained Variance
Endogenous Factor
Stigma Pity Expectancy Support
AIDS 22 I 31
Cancer 6 29 21
Drug abuse 9 51 15
Heart disease 1 31 11
Anorexia 4 51 6
Child abuse 9 36 29
Depression 16 14
Obesity 38 11
This is corroborated by the explained variance for the three endogenous vari-
ables pity, expectancy and support (Table 3). The model succeeded in explaining
a great deal of the variance of expectancy and support but much less so of pity.
This shows that the emotion of pity is not sufficiently predicted by controllability
and coping. Other factors, not under scrutiny here, must be responsible for the
variation in pity.
The stigma-specific path coefficients are displayed in Figures 1-8; Table 4

contains the decomposition of total effects into direct and indirect effects. Results
for each stigma will be described briefly. Coefficients above .21 are significant.
Table 4
Decomposition of Effects on Social Support Intention
Stigma Predictor Direct Indirect Total Effect
AIDS Control 0 21 21
Coping 0 14 14
Pity 55 0 55
Expectancy 7 0 7
Cancer Control 0 6 6
Coping 0 13 13
Pity 43 0 43
Expectancy 7 0 7
Drug abuse Control 0 6 6

Coping 0 29 29
Pity 14 0 14
Expectancy 40 0 40
Heart disease Control 0 3 3

Coping 0 16 16
Pity 19 0 19
Expectancy 26 0 26
Anorexia Control 0 0 0
Coping 0 17 17
Pity 14 0 14
Expectancy 24 0 24
Child abuse Control 0 12 12

Coping 0 25 25
Pity 37 0 37
Expectancy 31 0 31
Depression Control 0 4 4
Coping 0 14 14
Pity 14 0 14
Expectancy 34 0 34
Obesity Control 0 4 4
Coping 0 18 18
Pity 14 0 14
Expectancy 30 0 30
In the case of AlDS, a substantial causal path leads from controllability to

pity (p = -.39) and another from pity to support (p = .55). Coping has a somewhat
lower impact on pity (p = .25). Expectancy does not playa role: it is predicted
neither by controllability nor by coping, and it does not predict support. Since
AIDS is a terminal disease, it is not surprising not to find a large variation in out-
come expectancy. Pity appears to be the appropriate emotional reaction which
facilitates the likelihood to extend support (see Figure t and Table 4) .
-.39
Control Pity
.55
.25
Support
.06
.07
Coping Expect
.08
Figure I Pity and expectancy as mediators between controllability and coping and
social support in the AIDS scenario.
For cancer, pity was again the best predictor of support (p = .43), whereas
expectancy failed to contribute anything (p = .07). But the antecedents were dif-
ferent; controllability had no significant impact on pity or expectancy, whereas
coping had a strong path to expectancy (p = .52) and a moderate one to pity (p =
.22). Although cancer can be a terminal disease in many cases, there are better
survival chances for those who comply with treatment. This explains the associa-
tion between coping and expectancy, but, surprisingly, there was little effect on
support intentions that were based more on pity (see Figure 2 and Table 4) .
-.12
Control Pity
.43
.22
Support
.13
.07
Coping Expect
.52
Figure 2 Pity and expectancy as mediators between controllability and coping and
social support in the cancer scenario.
In case of drug abuse, variations in controllability elicited no effects reac-

tions but coping did. A strong path from coping to expectancy emerged (p .70), =
Expectancies and Help lntentions 77
accompanied by another strong path from expectancy to support (p = AD) making

this the major pathway to help intentions. A minor pathway was added from cop-
ing through pity (p = .25, P = .13). Drug abuse is a rather unstable condition and
appears to be modifiable. Whether one is ready to support a drug user mainly
depends on the likelihood of perceived change based on his or her coping efforts.
no matter how the problem was originally caused (see Figure 3 and Table 4) .
-.15
Control Pity
.13
.25
Support
.09
.40
Coping Expect
.70
social support in the drug abuse scenario.
In case of heart disease, there was no effect of controllability, and pity also
had no significant relationships (see Figure 4 and Table 4. The only pathway to
support led from coping via expectancy (p = .56, p = .26). Heart disease is inter-
preted as a modifiable condition that varies with one's health behavior such as
nutrition, exercise, and relaxation. The origin of this condition seems to be unim-
portant for a decision to help the patient.
-.04
Control Pity
.19
.06
Support
.07
.26
Coping Expect
.56
social support in the heart disease scenario.
Anorexia nervosa can also be regarded as an unstable condition where active

coping makes a difference. Controllability had no influence but coping deter-
mined expectancy (p = .70) and pity (p = .20) (see Figure 5 and Table 4).
Anorexia is considered a highly modifiable condition. If a patient copes well it

will vanish, no matter how controllable the origin was .
-.01
Control Pity
.14
.20
Support
.12
.24
Coping Expect
.70
social support in the anorexia scenario.
A different picture emerged for child abuse. Both direct effects on support
were almost equal, with pity (e = .37) and expectancy (e = .31) accounting for a
similar amount of variation in support. The key antecedent factor, however, was
coping which was closely related to expectancy (p = .60). Compared to drug
abuse, child abuse is not a health-compromising behavior but more a socially
deviant act that elicits emotions such as either outrage or pity towards the actor,
the latter emotion only if there was not much control over the behavior (see
Figure 6 and Table 4).
-.25
Control Pity
.37
.18
Support
.10
.31
Coping Expect
.60
social support in the child abuse scenario.
In case of depression, the predictors controllability and pity turned out to be

irrelevant, whereas expectancy had an influence on support (p = .34), based on
the coping efforts of the target person (p = .38). This clearly documents that an
active contribution on behalf of the mental health patient is required in order to
make the condition look changeable, so that support would not be in vain. Only
expectancy had an effect on support (see Figure 7 and Table 4).
Finally, in case of obesity, an almost identical result emerged. Again, control-

lability and pity were negligible factors but expectancy (p = .30), based on
coping (p = .61), made the difference. Obesity is an unstable condition, and those
who do not counteract their problem cannot count on help from others. Only
active coping efforts elicit expectancy which in tum trigger readiness for social
support (see Figure 8 and Table 4).
-.01
Control Pity
.14
.08
Support
.12
.34
Coping Expect
.38
social support in the depression scenario.
-.09
Control Pity
.14
.04
Support
.09
.30
Coping Expect
.61
Figure 8 Pity and expectancy as mediators between controllability and coping and social
support in the obesity scenario.
In sum, in five of the eight stigmas, outcome expectancy was the main pre-
dictor of support intention. These five were drug abuse, heart disease, anorexia,
depression, and obesity. The two terminal diseases, AIDS and cancer, differed
from the majority by their conspicuous pathway from pity to support intent. In
these two cases, one's intention to help was almost exclusively based on pity. For
child abuse, a balanced influence of pity and expectancy emerged. Coping was a
stronger antecedent than controllability in seven out of eight cases. The exception
was AIDS. The overall picture corroborates the assumption that outcome
expectancy is a critical mediator between target coping and social support inten-
tion. From these results, whether one extends help or not is primarily dependent
on the expectancies aroused by the victim characteristics, and particularly the
person's way of coping.
Discussion
Each of the eight stigmas was examined in separate path analyses with
respect to the two experimental factors, controllability and coping as antecedents,
and pity and expectancy as mediators. The model fit the data and expectancies
were a major direct source of support variation. Pity was a direct predictor of
social support only in three specific contexts. It is noteworthy that there was a
high degree of variation between the eight stigmas, indicating that the specific
circumstances decide whether the willingness to help is primarily based on either
pity or expectancy. In terminal diseases such as AIDS or cancer, pity appeared to
be more influential than expectancy, whereas for unstable health conditions such
as drug abuse, anorexia or obesity the coping-expectancy-support link was
obvious. It might be, therefore, that the perceived stability of a condition is a crit-
ical underlying dimension that affects judgments of help. Controllability was less
influential compared to coping which, in turn, partly determined expectancy. The
most conspicuous pathway led from coping via expectancy to support intent.
STUDY II
In the first experiment, the expected improvement of the target's condition
was one of the mediators under investigation. In the second experiment, the
attention was shifted to a support provider characteristic to address the question
of whether the perception of one's ability to help would make it more likely that
a support intention occurs. In other words, self-efficacy expectancy, one's per-
ceived personal capability of extending effective support, was the focus. It was
hypothesized that self-efficacy expectancy played the same role as a mediator
that outcome expectancy did in the first experiment. 2
Method
Design. The path-analytic model was the same as in the first study but there
were some differences in the experimental manipulations and in the measures in-
volved. Only one problem situation was selected, a sexual assault scenario, that
was varied with respect to controllability and coping. A rape victim in the
uncontrollable condition was described as a student who studied one night at the
library and was raped on the way to her car by a stranger. In the condition
designed to seem slightly more controllable she was described as someone who
attended a party where she drank too much and flirted with the males; when she
was taken home by one of them, she invited him up to her apartment and was
raped. The adaptively coping victim was characterized as one who was trying
hard to go on with her life after the assault, having joined a support group and
seeing a counselor each week. The maladaptively coping victim did not try to
2 Study II was conducted by Grace Woo, Christine Dunkel-Schetter, and Ralf

Schwarzer.
overcome her problem situation. She had withdrawn from friends and did not eat;
she also refused to attend a support group meeting and to see a counselor.
The experiment was arranged as a 2 x 2 between-subjects design; 70 under-
graduate students responded to the vignette randomJy assigned to one of four
conditions. There were 55 males and only 15 females, but their distribution over
the four cells was about equal, with cell sizes of 19, 17, 18 and 16.
Measures. Pity, self-efficacy expectancy and support intentions were the
dependent variables used in this report. All were rated on a 5-point scale. Pity
was assessed by four adjectives as part of a checklist, namely empathy, sym-
pathy, pity, and compassion. Emotional support intent was measured by four
items such as "Would you be willing to try to console and reassure your friend
when she is upset?" and "Would you spend time listening to her emotional reac-
tions to the assault?" Tangible support intent was measured by six items such as
"Would you be willing to offer her help with her school work if she needed it?"
and "Would you lend her money to see a therapist?" Self-efficacy expectancy
was measured by a newly developed 10-item scale that was employed for the first
time. Its psychometric properties were satisfactory with an average item-total
correlation of .55 and an internal consistency of Cronbach's alpha = .85. The
items were worded in the following way:
I. I possess the necessary social skills to alleviate the distress of a sexual assault
victim.
2. It is easy for me to comfort someone in distress.
3. I am capable of providing the appropriate resources for a rape victim.
4. It is difficult for me to communicate empathic understanding. (-)
5. I could make someone feel better no matter how depressed she is.
6. When it comes to comforting someone, I feel awkward. (-)
7. I am not sensitive enough to meet the support needs of a sufferer. (-)
8. I do not trust my skills to communicate in a beneficial way with a sexual assault
victim. (-)
9. I am not the kind of person who can meet the emotional needs of others who are in a
crisis. (-)
10. I have sufficient communication skills to cheer up someone who is experiencing
stress.
Results
A structural model was specified with the two experimental factors as ante-
cedents, and with pity, self-efficacy expectancy, and support as the dependent
variables. In contrast to the previous study, this is a multiple indicator model. The
three endogenous variables were specified with two indicators each. The four
pity items were divided into two sets (each pity indicator had two items); support
was specified by the emotional support scale as well as the tangible support scale,
and the two self-efficacy indicators were two 5-item subsets of the instrument
described above. The results of the LlSREL analysis are depicted in Figure 9.
.86 .70
-.12 Pity
Control
44
.23
Support
.06
.90 80
Self-eff. .36
Coping Expectancy
.25
.62 95
social support in the rape scenario.
The fit of the model was chi-square = 17.4 (15 df, p = .295) with a chi-
square/df ratio of 1.16. Goodness of fit was GFI = .94 and adjusted goodness of
fit AGFI = .87. The root mean square residual was RMSR = .09. Although the
latter two indices fall short of the usual requirement, the overall fit can be regard-
ed as satisfactory, based on the other indices. The explained variance for social
support was 34%, which is quite good. whereas those for pity and for self-
efficacy expectancy were low (7% each). Decomposing the effects on support led
=
to substantial direct effects for pity (e .44) and for self-efficacy expectancy (e =
.36), and to smaller indirect effects for controllability (e = .03). and for coping
(e = .17). Pity and expectancy were very good predictors of support intent, but
the underlying experimental factors (control, coping) were of lesser influence.
Discussion
The second experiment has replicated the general causal model leading from
victim characteristics to support provider emotional reactions or cognitions,
resulting in an intention formation. Pity emerged here as the strongest predictor
of support, but self-efficacy expectancy also contributed substantially. Controlla-
bility turned out to be negligible, whereas coping exerted a weak. but statistically
significant. influence on pity and self-efficacy.
However. it is difficult to construe a sexual assault as controllable, and the
two conditions differed in ratings of controllability only by one point, although
significantly. In addition. the rape scenario is quite different from the eight
stigma scenarios described in Study I. There is no disease or bodily condition
involved but a single violent act caused by an external agent. An assault is likely
to be viewed generally as less controllable than other social stigmas such as
obesity or drug abuse. The degree of controllability only varied in the study from
uncontrollable to somewhat controllable; there was not really a "controllable"
experimental condition. Adverse chance events seem especially likely to trigger
pity, whereupon the victim is not blamed.
GENERAL DISCUSSION
The present findings from Studies I and II are based on hypothetical
scenarios with students. Therefore, the results can be generalized neither to actual
helping situations nor to other populations. This procedure also has some inher-
ent limitations in that respondents may be unable to judge accurately their affec-
tive reactions and whether they would or would not offer help to particular indi-
viduals. In addition, some key variables that affect emotion and social support
certainly are excluded from the manipulated factors. However, as noted by
Cooper (1976), "when looked at from the point of view of generating hypotheses,
finding new leads, and initiating models of behavior, lrole playing I may be the
rbestl method" (p. 605). In addition, in the investigations presented here and by
Weiner et al. (1988), the stimulus configurations examined could not be found
without overwhelming difficulty in field research, with the consequences that
variables would be confounded. Finally, prior research has suggested that role-
enactment strategies in the study of help-giving have yielded data comparable
with observations of actual behavior (see review in Weiner, 1986). For these
reasons, and particularly in light of the relatively recent growth of the study of
social support, we used a hypothetical scenario method. Research must extend
theoretical and experimental analyses within the current framework before
applying these research questions to real-life situations.
The present studies have underscored the notion of emotional and cognitive
mediators in the process of forming behavioral intentions. When dealing with
victims of life events including medical patients, the likelihood of mobilizing
help is dependent on a number of recipient and provider characteristics (Dunkel-
Schetter & Skokan, 1990). The controllability of the cause of the problem ap-
pears to playa role in the determination of help. Moreover, the changeability or
instability of the problem as reflected in coping efforts seems to elicit positive
expectancies in the observer and motivation to help. Such efforts may create both
a sense that the situation can be improved and a belief that one can effectively as-
sist the victim. Thus, outcome expectancy as well as self-efficacy expectancy are
useful cognitive mediators. They are part of a mechanism that governs the trans-
lation of thought into action. Both studies have dealt with one of these cognitions
exclusively, and it would be worthwhile to integrate both concepts into one
empirical framework in a subsequent study.
One conclusion of the first experiment concerns the specificity of the result
pattern to individual situations. To what degree pity or expectancy mediate
reCIpIent characteristics and support intent depended on the particular

circumstances, i.e., the stigma chosen and, probably, the unique wording of the
vignettes. In the second experiment, there was only one context provided, namely
the rape scenario. Therefore, it remains unclear, as to whether these
circumstances have affected the results. It could be, for example, that for a
divorce or an accident, completely different path coefficients would emerge. The
evidence for self-efficacy expectancy as a mediator is limited to the context
chosen, and further research should make use of a number of different problem
domains.
There are underlying similarities, however, between the selected problems
that may suggest a common pattern of reactions to victims. For AIDS, cancer,
and rape, the emotion of pity appears to be a stronger mediator than expectancy.
These problems are loss/harm situations, whereas contexts such as anorexia,
obesity, drug abuse, child abuse, depression and heart disease are more like
threats (see Dunkel-Schetter et aI., 1991; Hobfoll, 1988; Lazarus, 1990; Lazarus
& Folkman, 1984). Different stress appraisals may determine the amount of pity
and specific expectancies in potential support providers. If a victim is severely
harmed or if the physical integrity of a victim is lost, then pity prevails; if, how-
ever, an on-going risky or threatening behavior is the topic, it is seen as more un-
stable and modifiable and, therefore, gives rise to a greater role for expectancies.
Expectancies can be pessimistic or optimistic. Pessimism undermines the
motivation to help because the investment of further support efforts appears to be
wasted; optimism, however, assumes that the victim will be responsive to future
support attempts and thereby render them worthwhile. Optimism, as a psycholog-
ical construct, has been defined as "generalized outcome expectancies" (Scheier
& Carver, 1985, 1987). This construct has recently become one of the key issues
in research on stress, coping, and mental health as well as physical health
(Scheier et aI., 1989; Seligman, 1991). The present studies have underscored the
role of situation-specific outcome expectancies and self-efficacy expectancies
after Bandura (1977, 1991). Further research should address the notion of specifi-
city versus generality of expectancy, with dispositional optimism being one
example of a more general construct. Jerusalem and Schwarzer (this volume)
have developed a global self-efficacy scale that has demonstrated high predictive
and construct validity in several field studies. Although specific measures are
preferred in clinical intervention studies of behavioral change, there might be an
advantage to global measures in other research domaiils.
Although the present studies have provided preliminary evidence for the role
of expectancies as mediators in the helping process, it remains undetermined how
outcome expectancy and self-efficacy expectancy are interrelated. Each experi-
ment has dealt with only one of these cognitions but failed to account for their
joint influence. It would be premature to conclude from the above findings that
outcome expectancy exerts a stronger influence on support intent than self-
efficacy expectancy. There might be a causal order among the two. For example,
it might be that a support provider does not scrutinize her helping capability
unless being faced with a target's condition that is improving or one that is, at
least, modifiable. A third variable could be critical here, namely one's personal
experience with (a) crisis situations that require support, and with (b) the effec-
tiveness of one's previous helping attempts (Dunkel-Schetter & Skokan, 1990).
Self-efficacy expectancy is shaped by context-specific mastery experiences,
among others, and therefore it would be necessary to investigate expectancies
jointly with an assessment of previous help experience.
REFERENCES
Ajzen, I. (1988). Attitudes, personality. and behavior. Milton Keynes: Open University
Press.
Amato, P. R. Ho, R, & Partidge, S. (1984). Responsibility attribution and helping
behavior in the Ash Wednesday bushfires. Australian Journal of Psychology, 36,
191-203.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change.
Psychological Review, 84, 191-215.
Bandura, A. (1986). Social foundations of thought and action. Englewood Cliffs, NJ:
Prentice-Hall.
Bandura, A. (1991). Self-efficacy conception of anxiety. [n R. Schwarzer &
R Wicklund (Eds.), Anxiety and selflocused attention (pp. 89-110). London:
Harwood.
Barbee, A. P. (1990). Interactive coping: The cheering-up process in close relationships.
In S. Duck with R Silver (Ed.), Social support and personal relationships (pp. 46-
65). London: Sage.
Batson, C. D. (1990). How social an animal? American Psychologist, 45,336-346.
Batson, C. D., O'Quinn, K., Fultz, J., Vanderplas, M., & Isen, A. (1983). Influence of
self-report distress and empathy on egoistic versus altruistic motivation to help.
Bennett-Herbert, T., & Dunkel-Schetter, C. (in press). Negative social reactions to
victims: An overview of responses and their determinants. In L. Montada, S. H.
Filipp, & M. J. Lerner (Eds.), Life crises and experiences of loss in adulthood.
Hillsdale, NJ: Erlbaum.
Bentler, P. M. (1980). Multivariate analysis with latent variables: Causal modeling.
Berkowitz, L. (1987). Mood, self-awareness, and willingness to help. Journal of
Personality and Social Psychology, 52,721-729.
Betancourt, H. (1990). An attribution-empathy model of helping behavior: Behavioral
intentions and judgments of help-giving. Personality and Social Psychology Bulletin,
16,573-591.
Borgida, E., Simmons, R. G., Conner, C., & Lombard, K. (1990). The Minnesota Living
Donor Studies: Implications for organ procurement. In J. Shanteau &. R. Harris
(Eds.), Psychological research on organ donation (pp. 108-121). Washington, DC:
American Psychological Association.
Brickman, P., Rabinowitz, P. c., Karuza, J. Jr., Coates, D., Cohn, E., & Kidder, L.
(1982). Models of helping and coping. American Psychologist, 37, 368-384.
Carlson, M., & Miller, N. (1987). Explanation of the relation between negative mood
and helping. Psychological Bulletin, 102,72-90.
Cialdini, R., Darby, B., & Vincent, J. (1973). Transgression and altruism: A case for
hedonism. Journal of Experimental Social Psychology, 9, 502-516.
Clary, E. G., & Orenstein, L. (1991). The amount and effectiveness of help: The
relationship of motives and abilities to helping behavior. Personality and Social
Psychology Bulletin, 17,58-64 .
. Cooper, J. (1976). Deception and role playing. American Psychologist, 31, 605-610.
Dalbert, c., Montada, L., & Schmitt, M. (1988). Intention and ability as predictors of
change in adult daughters' prosocial behavior towards their mothers. Verantwortung.
Gerechtigkeit. Moral, 46, 1-17.
Dejong, W. (1980). The stigma of obesity: The consequences of naive assumptions
concerning the causes of physical deviance. Journal of Health and Social Behavior,
21,75-87.
Dovidio, J. F. (1984). Helping behavior and altruism: An empirical and conceptual
overview. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol.
17, pp. 362-428). New York: Academic Press.
Dunkel-Schetter, c., & Bennett, T. L. (1990). Differentiating the cognitive and behav-
ioral aspects of social support. In I. G. Sarason, B. R. Sarason, & G. R. Pierce (Eds.),
Social support: An interactional view (pp. 267-296). New York: Wiley.
Dunkel-Schetter. C., Blasband, D. E., Feinstein, L. G., & Bennett, T. L. (1991). Ele-
ments of supportive social interactions: When are support attempts effective? In
S. Spacapan & S. Oskamp (Eds.), Helping and being helped in the real world (pp. 83-
114). Newbury Park, CA: Sage.
Dunkel-Schetter, C., Folkman, S., & Lazarus. R. S. (1987). Correlates of social support
receipt. Journal of Personality and Social Psychology, 53. 71-80.
Dunkel-Schetter, c., & Skokan, L. A. (1990). Determinants of social support provision
in personal relationships. Journal oj Social and Personal Relationships. 7, 437-450.
Dunkel-Schetter, c., & Wortman, C. B. (1981). Dilemmas of social support: Parallels
between victimization and aging. In I. B. Kiesler, 1. N. Morgan, & V. K. Oppen-
heimer (Eds.), Aging: Social change (pp. 349- 381). New York: Academic Press.
Dunkel-Schetter, C., & Wortman, C. B. (1982). The interpersonal dynamics of cancer:
Problems in social relationships and their impact on the patient. In H. S. Friedman &
M. R. DiMatteo (Eds.). Interpersonal issues in health care (pp. 69-100). New York:
Academic Press.
Eisenberg, N., & Miller. P. A. (1987). The relation of empathy to prosocial and related
behaviors. Psychological Bulletin, 101,91-119.
Fishbein, M., & Ajzen, 1. (1975). Belief, attitude. intention. and behavior: An introduc-
tion to theory and research. Reading, MA: Addison-Wesley.
Goranson, R. E., & Berkowitz, L. (1966). Reciprocity and responsibility reactions to
prior help. Journal of Personality and Social Psychology, 3,227-232.
Gurtman, M. B. (1986). Depression and the response of others: Re-evaluating the eval-
uation. Journal of Abnormal Psychology, 95, 99-101.
Hobfoll, S. E. (1988). The ecology of stress. Washington, DC: Hemisphere.
Hoffman, M. L. (1975). Developmental synthesis of affect and cognition and its impli-
cations for altruistic motivation. Developmental Psychology, II, 607-622.
Hoffman, M. L. {I 982). Development of pro social motivation: Empathy and guilt. In
N. Eisenberg (Ed.), Development oj prosocial behavior (pp. 281-313). New York:
Academic Press.
Joreskog, K., & Sorbom, D. (l988). LlSREL VII: A guide to the program and applica-
tions. Mooresville, IN: Scientific Software.
lones, E. E., Farina, A., Hastorf, A. H., Markus, H., Miller, D., & Scott, R. A. (1984).
Social stigma. The psychology of marked relationships. New York: Freeman.
lung, l. (1988). Social support providers: Why do they help? Basic and Applied
Karasawa, K. (1991). The effects of onset and offset responsibility on affects and help-
ing judgments. Journal of Applied Social Psychology, 21,482-499.
Lazarus, R. S. (1990). Theory-based stress measurement. Psychological Inquiry, 1,3-13.
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York:
Springer.
Notarius, C. I., & Herrick, L. R. (1988). Listener response strategies to a distressed
other. Journal of Social and Personal Relationships, 5,97- 108.
Reisenzein, R. (1986). A structural equation analysis of Weiner's attribution-affect
model of helping behavior. Journal of Personality and Social Psychology, 50, 1123-
1133.
Scheier, M. F., & Carver, C. S. (1985). Optimism, coping, and health: Assessment and
implications of generalized outcome expectancies. Health Psychology, 4,219-247.
Scheier, M. F., & Carver, C. S. (1987). Dispositional optimism and physical well-being:
The influence of generalized outcome expectancies on health. Journal of Personality,
55,169-210.
Scheier, M. F., Matthews, K. A., Owens, 1., Magovern, G. 1. Sr., Lefebre, R. c., Abbott,
R. A., & Carver, C. S. (1989). Dispositional optimism and recovery from coronary
artery bypass surgery: The beneficial effects on physical and psychological well-
being. Journal of Personality and Social Psychology.
Schmidt, G., & Weiner, B. (1988). An attribution-affect-action theory of motivated
behavior: Replications examining judgments of help-giving. Personality and Social
Psychology Bulletin, 14,610-621.
Schwarzer, R., & Leppin, A. (1991). Social support and health: A theoretical and
empirical overview. Journal of Social and Personal Relationships, 8,99-127.
Schwarzer, R., & Weiner, B. (\ 990). Die Wirkung von Kontrollierbarkeit und Bewalti-
gungsverhalten auf Emotionen und soziale Unterstiitzung rThe effect of
controllability and coping on emotions and social supportl. Zeitsrhrift fur Sozialpsy·
chologie, 2/,118-125.
Schwarzer, R., & Weiner, B. (1991). Stigma controllability and coping as predictors of
emotions and social support. Journal of Social and Personal Relationships, 8, 133-
140.
Seligman, M. E. P. (1991). Learned optimism. New York: Knopf.
Shumaker, S. A., & Brownell, A. (1984). Toward a theory of social support: Closing
conceptual gaps. Journal of Social Issues, 40, 11-36.
Silver, R. c., Wortman, C. 8., & Crofton, C. (1990). The role of coping in support pro-
vision: The self-presentational dilemma of victims of life crises. In I. G. Sarason,
8. R. Sarason, & G. R. Pierce (Eds.), Social support: An interactional view (pp. 397-
426). New York: Wiley.
Skokan, L. A. (1990). Motivational determinants of social support provision.
Unpublished doctoral dissertation. University of California: Los Angeles.
Weiner, B. (1985). An attributional theory of achievement motivation and emotion.
Weiner, B. (1986). An attributional theory of motivation and emotion. Berlin: Springer.
Weiner, B., Perry, R. P., & Magnusson, l. (1988). An attributional analysis of reactions
to stigmas. Journal of Personality and Social Psychology, 55(5) 738-748.
References
I Self-Efficacy and Human Functioning
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes

and predicting social behav
Ajzen, I., & Madden, T. 1. (1986). Prediction of

goal-directed behavior: Attitudes,
intentions, and perceived behavioral control. Journal of

Experimental Social
Alden, L. (1986). Self-efficacy and causal attributions

for social feedback. Journal of
Atkinson, J. W. (1964). An introduction to motivation.

Princeton, NJ: Van Nostrand.
Bandura, A. (1986). Social foundations of thought and

action: A social cognitive theory.
Bandura, A. (1988). Perceived self-efficacy: Exercise of

control through self-belief. In
J. P. Dauwalder, M. Perrez, & V. Hobi (Eds.), Annual series

of European research in
behavior therapy (Vol. 2, pp. 27-59). Lisse, The

Netherlands: Swets & Zeitlinger.
Bandura, A. (1991). Self-regulation of motivation through

anticipatory and self-regula
tory mechanisms. In R. A. Dienstbier (Ed.), Perspectives

on motivation: Nebraska
symposium on motivation (Vol. 38, pp. 69-164). Lincoln:

University of Nebraska
Press.
Bandura, A., & Abrams, K. (1986). Self-regulatory
mechanisms in motivating. apathet
ic. and despondent reactions to unfulfilled standards.

Unpublished manuscript,
Stanford University, Stanford.
Bandura, A., & Adams, N. E. (1977). Analysis of

self-efficacy theory of behavioral
change. Cognitive Therapy and Research, 1,287-308.
Bandura, A .• & Cervone, D. (1983). Self-evaluative and

self-efficacy mechanisms gov
erning the motivational effects of goal systems. Journal

of Personality and Social
Psychology, 45, 1017-1028.
Bandura, A., & Cervone, D. (1986). Differential engagement

of self-reactive influences
in cognitive motivation. Organizational Behavior and Human

Decision Processes,
38,92-113.
Bandura. A., Cioffi, D., Taylor, C. B., & Brouillard. M. E.

(1988). Perceived self-effi
cacy in coping with cognitive stressors and opioid

activation. Journal of Personality
Bandura. A., & Jourden, F. 1. (1991). Self-regulatory

mechanisms governing the im
pact of social-comparison on complex decision making.

Journal of Personality and
Social Psychology, 60. 941-951.
Bandura, A .• Reese, L., & Adams, N. E. (1982).

Microanalysis of action and fear arous
al as a function of differential levels of perceived

self-efficacy. Journal of Personality
Bandura, A., Taylor, C. B., Williams, S. L, Mefford, I. N.,

& Barchas, 1. D. (1985).
Catecholamine secretion as a function of perceived coping

self-efficacy. Journal of
ConSUlting and Clinical Psychology, 53,406-414.
Bandura, A., & Wood, R. E. (1989). Effect of perceived

controllability and perform
ance standards on self-regulation of complex

decision-making. Journal of Personality
Bandura, M. M., & Dweck, C. S. (1988). The relationship of

conceptions of intelli
gence and achievement goals to achievement-related

cognition. affect and behavior.
Unpublished manuscript.
Barling, J., & Abel, M. (1983). Self-efficacy beliefs and

tennis performance. Cognitive
Barling, J., & Beattie, R. (1983). Self-efficacy beliefs

and sales performance. Journal
of Organizational Behavior Management, 5, 41-51.
Beach, L. R., Bames, V. E., & Christensen-Szalanski, 1. 1.

1. (1986). Beyond heuristics and biases: A contingency
model of jUdgmental forecasting. Journal of Forecasting,
5,143-157.
Beck, A. T., Emery, G., & Greenberg, R. L. (1985). Anxiety

disorders and phobias.
Beck, K. H., & Lund, A. K. (1981). The effects of health

threat seriousness and person
al efficacy upon intentions and behavior. Journal of
Applied Social Psychology, II,
401-415.
Becker, L. 1. (1978). Joint effect of feedback and goal

setting on performance: A field
study of residential energy conservation. Journal of

Betz, N. E., & Hackett, G. (1986). Applications of

self-efficacy theory to understanding
career choice behavior. Journal of Social and Clinical

Blustein, D. L. (1989). The role of goal instability and

career self-efficacy in the career
exploration process. Journal of Vocational Behavior, 35,

194-203.
Bower, G. H., (1983). Affect and cognition. Philosophical

Transactions of the Royal
Society of London (Series B), 302,387-402.
Carver, C. S., & Scheier, M. F. (1981). Attention and

self-regulation: A control-theory
approach to human behavior. New York: Springer-Verlag.
Cervone, D. (1989). Effects of envisioning future

activities on self-efficacy judgments
and motivation: An availability heuristic interpretation.

Cognitive Therapy and
Research, 13,247-261.
Cervone, D., Jiwani, N., & Wood, R. (1991). Goal-setting

and the differential influence
of self-regulatory processes on complex decision-making

performance. Journal of
Personality and Social Psychology, 6/,257-266.

Cervone, D., & Peake, P. K. (1986). Anchoring, efficacy,
and action: The influence of
judgmental heuristics on self-efficacy judgments and

behavior. Journal of Personality
Churchill, A. C. (1990). Metacognitive self-efficacy and

intrusive thought. Unpublished
doctoral dissertation. University of Melbourne, Australia.
Churchill, A. c., & McMurray, N. E. (1990). Self-efficacy

and unpleasant intrusive
thought. Submitted for publication.
Coe, C. L., & Levine, S. (1991, in press).

Psychoimmunology: An old idea whose time
has come. In P. R. Barchas (Ed.), Social physiology of

social relations. Oxford:
Collins, 1. L. (1982, March). Self-efficacy and ability in

achievement behavior. Paper
presented at the annual meeting of the American

Educational Research Association,
N.Y.
Corbin, C. (1972). Mental practice. In W. Morgan (Ed.),

Ergogenic aids and muscular
performance (pp. 93-118). New York: Academic Press.
Cutrona, C. E., & Troutman, B. R. (1986). Social support,

infant temperament, and
parenting self-efficacy: A mediational model of postpartum

depression. Child
Davis, F. W., & Yates, B. T. (1982). Self-efficacy

expectancies versus outcome ex
pectancies as determinants of performance deficits and
depressive affect. Cognitive
De Vries, H., Dijkstra, M., & Kuhlman, P. (1988).

Self-efficacy: The third factor be
sides attitude and subjective norm as a predictor of

behavioural intentions. Health
Dweck, C. S., & Leggett, E. L. (1988). A social-cognitive

approach to motivation and
personality. Psychological Review, 95, 256-273.
Dzewaltowski, D. A. (1989). Towards a model of exercise

motivation. Journal 0/ Sport
and Exercise Psychology, / /,251-269.
Dzewaltowski, D. A., Noble, J. M., & Shaw, 1. M. (1991).

Physical activity participa
tion: Social cognitive theory versus the theories of

reasoned action and planned
behavior. Journal of Sport and Exercise Psychology, 12,

388-405.
Feather, N. T. (Ed.) (1982). Expectations and actions:

Expectancy-value models in
psychology. Hillsdale, NJ: Erlbaum.
Feltz, D. L., & Landers, D. M. (1983). Effects of mental

practice on motor skill learn
ing and performance: A meta-analysis. Journal of Sport

Godding, P. R., & Glasgow, R. E. (1985). Self-efficacy and

outcome expectations as
predictors of controlled smoking status. Cognitive Therapy

and Research, 9, 583-590.
Hackett, G., & Betz, N. E. (1981). A self-efficacy

approach to the career development
of women. Journal o/Vocational Behavior, /8,326-339.
Hackett, G., & Betz, N. E. (1984). Mathematics performance,

mathematics self-efficacy, and the prediction of
science-based college majors. Unpublished manuscript,
University of California, Santa Barbara.
Hogarth, R. (1981). Beyond discrete biases: Functional and

dysfunctional aspects of
judgmental heuristics. Psychological Bulletin, 90, 197-217.
Holahan, C. K., & Holahan, C. J. (1987a). Self-efficacy,

social support, and depression in aging: A longitudinal
analysis. Journal of Gerontology, 42,65-68.
Holahan, C. K., & Holahan, C. 1. (1987b). Life stress,

hassles, and self-efficacy in aging: A replication and
extension. Journal of Applied Social Psychology, 17, 574
592.
Holroyd, K. A., Penzien, D. 8., Hursey, K. G., Tobin, D.

L., Rogers, L., Holm, 1. E., Marcille, P. J., Hall, 1.
R., & Chila, A. G. (1984). Change mechanisms in EMG
biofeedback training: Cognitive changes underlying
improvements in tension headache. Journal of Consulting
and Clinical Psychology, 52, 1039-1053.
Isen, A. M. (1987). Positive affect, cognitive processes,

and social behavior. In L. Berkowitz (Ed.), Advances in
experimental social psychology (Vol. 20, pp. 203253). New
York: Academic Press.
Jacobs, B., Prentice-Dunn, S., & Rogers, R. W. (1984).

Understanding persistence: An interface of control theory
and self-efficacy theory. Basic and Applied Social
10urden, F. 1., Bandura, A., & Banfield, 1. T. (1991, in

press). Impact of conceptions of ability on
self-regulatory mechanisms and skill development. Journal
of Sport and Exercise Psychology.
Kanfer, R., & Zeiss, A. M. (1983). Depression,

interpersonal standard-setting, and
judgments of self-efficacy. Journal of Abnormal

Kavanagh, D. 1. (1983). Mood and self-efficacy.

Unpublished doctoral dissertation,
Kavanagh, D. J., & Bower, G. H. (1985). Mood and

self-efficacy: Impact of joy and sadness on perceived
capabilities. Cognitive Therapy and Research, 9, 507-525.
Kavanagh, D. 1., & Wilson, P. H. (1989). Prediction of

outcome with a group version of cognitive therapy for
depression. Behaviour Research and Therapy, 27,333-343.
Kazdin, A. E. (1978). Covert modeling-Therapeutic

application of imagined rehearsal. In 1. L. Singer & K. S.
Pope (Eds.) The power of human imagination: New
methods in psychotherapy. Emotions, personality, and

psychotherapy (pp. 255-278). New York: Plenum.
Kazdin, A. E. (1979). Imagery elaboration and self-efficacy

in the covert modeling
treatment of unassertive behavior. Journal of Consulting

and Clinical Psychology,
47, 725-733.
Kent, G. (1987). Self-efficacious control over reported

physiological, cognitive and
behavioural symptoms of dental anxiety. Behaviour Research

and Therapy, 25,341347.
Kent, G., & Gibbons, R. (1987). Self-efficacy and the

control of anxious cognitions.
Journal of Behavior Therapy & Experimental Psychiatry, 18,

33-40.
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal,

and coping. New York:
Springer.
Lee, C. (1984a). Accuracy of efficacy and outcome

expectations in predicting perform
ance in a simulated assertiveness task. Cognitive Therapy
Lee, C. (1984b). Efficacy expectations and outcome

expectations as predictors of per
formance in a snake-handling task. Cognitive Therapy and

Research, 8, 509-516.
Leland, E. 1. (1983). Self-efficacy and other variables as

they relate to precompetitive
anxiety among male interscholastic basketball players.

(Doctoral dissertation, Stan
ford University, 1983). Dissertation Abstracts

International, 44, 1376A.
Lent, R. W., Brown, S. D., & Larkin, K. C. (1987).

Comparison of three theoretically
derived variables in predicting career and academic

behavior: Self-efficacy, interest
congruence, and consequence thinking. Journal of Counseling

Psychology, 34, 293
298.
Lent, R. W., & Hackett, G. (1987). Career self-efficacy:

Empirical status and future di
rections. Journal of Vocational Behavior, 30,347-382.
Litt, M. D. (1988). Self-efficacy and perceived control:

Cognitive mediators of pain
tolerance. Journal of Personality and Social Psychology,

54, 149-160.
Locke, E. A., Frederick, E., Lee, c., & Bobko, P. (1984).

Effect of self-efficacy, goals,
and task strategies on task performance. Journal of

Applied Psychology, 69, 241-25l.
Locke, E. A., & Latham, G. P. (1990). A theory of goal

setting and task performance.

Lord, R. G., & Hanges, P. J. (1987). A control system
model of organizational motiva
tion: Theoretical development and applied implications.

Behavioral Science, 32,
161-178.
Maier, S. F., Laudenslager, M. L., & Ryan, S. M. (1985).

Stressor controllability, im
mune function, and endogenous opiates. In F. R. Brush & J.

B. Ovennier (Eds.),
Affect, conditioning, and cognition: Essays on the

determinants of behavior (pp. 183
Major, B., Cozzarelli, c., Sciacchitano, A. M., Cooper, M.

L., Testa, M., & Mueller,
P. M. (1990). Perceived social support, self-efficacy, and

adjustment to abortion.
McAuley, E. (1985). Modeling and self-efficacy: A test of

Bandura's model. Journal
of Sport Psychology, 7, 283-295.
McCaul, K. D., O'Neill, K., & Glasgow, R. E. (1988).

Predicting the performance of
dental hygiene behaviors: An examination of the Fishbein

and Ajzen model and self
efficacy expectations. Journal of Applied Social

Psychology, /8, 114-128.
Meichenbaum, D. H. (1977). Cognitive-behavior

modification: An integrative ap
proach. New York: Plenum Press.
Miller, G. A., Galanter, E., & Pribram, K. H. (1960). Plans

and the structure of behav
ior. New York: Holt.
Nicholls, J. G. (1984). Achievement motivation:

Conceptions of ability, subjective ex
perience, task choice, and performance. Psychological

Review, 91, 328-346.
Olioff, M., & Aboud, F. E. (1991). Predicting postpartum

dysphoria in primiparous mothers: Roles of perceived
parenting self-efficacy and self-esteem. Journal of
Cognitive Psychotherapy, 5,3-14.
Ozer, E., & Bandura, A. (1990). Mechanisms governing

empowerment effects: A selfefficacy analysis. Journal of
Personality and Social Psychology, 58, 472-486.
Peake, P. K., & Cervone, D. (1989). Sequence anchoring and

self-efficacy: Primacy effects in the consideration of
possibilities. Social Cognition, 7, 31-50.
Perry, D. G., Perry, L. C., & Rasmussen, P. (1986).

Cognitive social learning mediators of aggression. Child
Post-Kammer, P., & Smith, P. (1985). Sex differences in

career self-efficacy, conside
ration, and interests of eighth and ninth graders. Journal

of Counseling Psychology,
32,551-559.
Relich, J. D., Debus, R. L., & Walker, R. (1986). The

mediation role of attribution and self-efficacy variables
for treatment effects on achievement outcomes. Contemporary
Educational Psychology, J I, 195-216.
Rotter, J. B. (1954). Social learning and clinical

psychology. Englewood Cliffs, NJ: Prentice-Hall.
Salkovskis, P. M., & Harrison, J. (1984). Abnormal and

normal obsessions-a replica
tion. Behaviour Research and Therapy, 22,549-552.
Sarason, I. G. (1975). Anxiety and self-preoccupation. In

I. G. Sarason & c. D. Spiel
berger (Eds.), Stress and anxiety (Vol. 2, pp. 27-44).
Washington, DC: Hemisphere.
Schunk, D. H., & Cox, P. D. (1986). Strategy training and

attributional feedback with
learning disabled students. Journal of Educational

Schunk, D. H., & Gunn, T. P. (1986). Self-efficacy and

skill development: Influence
of task strategies and attributions. Journal of Educational

Research, 79, 238-244.
Schunk, D. H., & Rice, J. M. (1986). Extended attributional

feedback: Sequence ef
fects during remedial reading instruction. Journal of

Early Adolescence, 6, 55-66.
Schwarzer, R. (1992). Self-efficacy in the adoption and

maintenance of health behav
iors: A critical analysis of theoretical approaches and a

new model. In Schwarzer, R.
(Ed.), Self-efficacy: Thought control of action. New York:

Hemisphere (this
volume).
Shavit, Y., & Martin, F. C. (1987). Opiates, stress, and

immunity: Animal studies.
Siegel, R. G., Galassi, J. P., & Ware, W. B. (1985). A

comparison of two models for
predicting mathematics performance: Social learning versus

math aptitude-anxiety.
Journal of Counseling Psychology, 32,531-538.
Silver, W. S., Mitchell, T. R., & Gist, M. E. (1989). The

impact of self-efficacy on
causal attributions for successful and unsuccessful

peiformance. Unpublished manu
script, University of Washington, Seattle, Washington.
Snyder, M. (1987). Public appearances, private realities:

The psychology of self
monitoring. New York: W. H. Freeman.
Strang, H. R., Lawrence, E. C., & Fowler, P. C. (1978).

Effects of assigned goal level
and knowledge of results on arithmetic computation:

Laboratory study. Journal of
Taylor, M. S., Locke, E. A., Lee, c., & Gist, M. E.

(1984). Type A behavior and facul
ty research productivity: What are the mechanisms?

Organizational Behavior and
Human Peiformance, 34,402-418.
Weinberg, R. S., Gould, D., & Jackson, A. (1979).

Expectations and performance: An
empirical test of Bandura's self-efficacy theory. Journal

of Sport Psychology, J, 320
331.
Wheeler, K. G. (1983). Comparisons of self-efficacy and

expectancy models of occu
White, J. (1982). Rejection. Reading, MA: Addison-Wesley.
Wiedenfeld, S. A., O'Leary, A., Bandura, A., Brown, S.,

Levine, S., & Raska, K.
(1990). Impact of perceived self-efficacy in coping with

stressors on components of
the immune system. Journal of Personality and Social

Psychology, 59, 1082-1094.
Williams, S. L., Dooseman, G., & Kleifield, E. (1984).

Comparative power of guided
mastery and exposure treatments for intractable phobias.

Journal of Consulting and
Williams, S. L., Kinney, P. J., & Falbo, J. (1989).

Generalization of therapeutic changes
in agoraphobia: The role of perceived self-efficacy.

Journal of Consulting and Clini
cal Psychology, 57,436-442.
Williams, S. L., & Rappoport, A. (1983). Cognitive

treatment in the natural environ
ment for agoraphobics. Behavior Therapy, 14,299-313.
Williams, S. L., Turner, S. M., & Peer, D. F. (1985).

Guided mastery and performance
desensitization treatments for severe acrophobia. Journal

of ConSUlting and Clinical
Williams, S. L., & Watson, N. (1985). Perceived danger and

perceived self-efficacy as
cognitive mediators of acrophobic behavior. Behavior

Therapy, /6, 136-146.
Wood, R. E., & Bandura, A. (l989a). Impact of conceptions

of ability on self-regulatory mechanisms and complex
decision making. Journal of Personality and Social
Wood, R. E., & Bandura, A. (1989b). Social cognitive

theory of organizational management. Academy of
Management Review, 14,361-384. Author Notes
This chapter includes revised and expanded material from an

article published in
The Psychologist as an invited address at the annual

meeting of the British
Psychological Society, St. Andrews, Scotland, April 1989.
TWO DIMENSIONS OF PERCEIVED SELF -EFFICACY: COGNITIVE

CONTROL AND BEHAVIORAL COPING ABILITY William J. McCarthy
and Michael D. Newcomb Confirmatory factor analyses were
conducted to test the empirical justification for
distinguishing between perceptions of behavioral coping
ability and perceptions of cognitive control coping
ability for handling environmental challenges. Twenty-four
measures of perceived personal effectiveness were collected
from 739 young adults. including measures of perceived
ability to have a social impact. assertiveness. leadership
style. and dating competence. These items were submitted
to a hierarchical confirmatory factor analysis in a random
half of the sample. As expected. two empirically
we/l-justified second order factors were obtained
reflecting perceived cognitive control and behavioral
coping strategies. This factor structure was
cross-validated in the other half of the sample. and
separately for males and females. with all hypothesized
features confirmed. Literature on coping strategies, on sex
role differences and on self-efficacy predictors is cited
as support for distinguishing between perceived cognitive
control and perceived behavioral coping abilities.
Implications of this distinction for elucidating
developmental patterns of drug use and for improving
understanding of relapse in lifestyle change programs are
discussed.
Why do individuals rely primarily on intrapsychic coping in

some contexts and
behavioral coping in other contexts? Suppose that a

middle-aged woman whose
children have grown and left home is distressed to find

that her husband pays her
inadequate attention, responding perfunctorily to her

comments and showing
more enthusiasm for newspaper reading and TV watching than

for talking with
her. Her perceived ability to cope cognitively with this

situation may be high or
low. If it is high, then by merely reframing her thoughts,

she can palliate or eli
minate her distress, perhaps by thinking of evidence that
she is indeed an inter
esting person and that her husband pays her no attention

because he is preoccu
pied by his work. If it is low, then the distressful

observation that her husband
finds her boring will intrude on her thoughts unless the

situation changes. Similarly, her perceived ability to
cope behaviorally with the situation may
be high or low. If it is high, she may elect to leave him

or feel confident that she
can alter his behavior. If it is low, she will feel that

the situation is inescapable
and that all of her alternatives are less attractive than

the status quo. Perceptions of personal coping ability
have been related to a wide range of
health-related outcomes, including smoking cessation,

weight control, alcohol
abuse, exercise, and contraceptive behavior (e.g.,

Strecher, DeVellis, Becker, &
Rosenstock, 1986, and O'Leary, 1985). As individuals'

self-percepts of coping
ability increase, so does the probability of their

achieving self-set health goals. Between the
identification of an important self-relevant goal and the
ultimate
achievement of the goal are interposed challenges with

which the individual must
cope. These challenges may be primarily cognitive or

primarily behavioral in
nature. The coping behaviors appropriate for dealing with

these challenges have
been termed emotion-focused or problem-focused (e.g.,

Folkman & Lazarus,
1980). Emotion-focused coping includes such behaviors as

avoidance, intellec
tualization, isolation, suppression, and magical thinking.

Problem-focused cop
ing includes such behaviors as information-seeking,

cognitive problem-solving,
inhibition of action and direct action. The perception that

one can effectively
implement emotion-focused or problem-focused coping can be

termed perceived
cognitive control ability and perceived behavioral coping

ability, respectively. Other literature on coping has
promoted a distinction between behavioral and
cognitive ways of coping. For example, Pearlin and Schooler

(1978) discussed
three major categories of coping responses, two of which

involved cognitive
strategies to reduce or eliminate stress, whereas the third

concerned the active
manipulation of the environment. In her review, Taylor

(1986) identified four
types of control that mediated the effects of coping with

stressors, but concluded
that these four types of control could be reduced to two:

(a) changing thoughts
with respect to the stressor, and (b) taking some action

with respect to the stres
sor. If people differ in whether they rely primarily on

cognitive or behavioral
means of coping with a challenge, they also probably differ

in their perceived
ability to use either cognitive control or behavioral

strategies for coping with the
challenges. The following report seeks to confirm the

validity and usefulness of
distinguishing between perceived cognitive control and
behavioral coping ability
through confirmatory factor analysis of young adult data on

coping strategies and
through example. Although we find it useful to distinguish

between perceived cognitive control
ability and perceived behavioral coping ability, we note

that efficacious behavior
is rarely a function exclusively of only one of these.

Characteristics of the con
text (such as the amount of freedom individuals have to

change the environment)
and characteristics of the individual (such as age)

determine which type of per
ceived coping ability is the more important contributor to

self-perceived ability to
perform the desired behavior. In their study of adult

responses to 1,332 stress episodes, Folkman and
Lazarus (1980) noted that both problem-focused and

emotion-focused coping
were used to cope with 98% of the episodes. They also noted
that the importance
of the type of coping varied, however, with context and

according to the charac
teristics of the individuals. Cognitive strategies were

employed most frequently
in situations where the individual was relatively helpless

to bring about the
desired behavior by themselves, such as when recovering

from an illness.
Problem-focused strategies were employed more frequently in

work situations.
Folkman and Lazarus found that the importance of type of

coping varied with
gender, with men relying more heavily than women on

problem-focused coping
even when the context permitted only emotion-focused

coping. It should be not
ed, again, that distinguishing conceptually between

perceived cognitive control
and perceived behavioral coping ability should not imply

that we view these con
cepts as independent. Folkman and Lazarus (1980), in fact,

observed a mean cor
relation of .44 between emotion-focused and problem-focused

coping across
three different samples. This study was designed to

confirm the reasonableness of distinguishing
generically between perceived cognitive coping ability and

perceived behavioral
coping ability as separate components of self-efficacy.

The importance of this
distinction, if accepted, is that it would be likely to

stimulate more careful exami
nation of the contextual, temporal, and individual

influences on percepts of self
efficacy. Individuals may have similar perceptions

concerning their respective
abilities to accomplish their jobs or to effect major

lifestyle changes, but never
theless vary greatly in their perceived ability to cope

with the behavioral or cog
nitive challenges associated with accomplishing the desired

behavior. Moreover,
within individuals the relative importance of perceived

cognitive control and be
havioral coping ability may vary with time, age and the
individual's experience
with coping with the specific challenge. In settings such

as prisons, where inmates would find that a problem-focused
coping strategy is often inappropriate, individuals can

still vary in their ability to
cope with environmental stressors, depending on their

perceived ability to regu
late their thoughts. In the same vein, children's responses

to environmental chal
lenges often are limited to cognitive coping strategies

because their dependency
on adults and their immaturity are such as to obviate the

use of behavioral coping
strategies. A child who is sexually abused by an adult

relative, for instance, typi
cally relies on cognitive strategies to cope with the

situation, especially disasso
ciation, and later, amnesia (Courtois, 1988).

Contrariwise, the theoretical necessity for the separate
concept of a perceived
ability to cope behaviorally seems justified by attempts to

relate the use of coping
strategies to spontaneous major lifestyle change.

Spontaneous major lifestyle
changes, such as taking up jogging, reducing the percentage

of calories in one's
diet derived from fat, and adopting a child, put a premium

on behavioral coping
relative to cognitive control coping, because there are

simply too many changes
in day-to-day behaviors resulting from the major lifestyle

change in question to
be anticipated cognitively. It is reasonable to assume that

persons with strong
beliefs in their general ability to change their social and

physical environment
will be more likely to embark on major voluntary lifestyle

changes than are per
sons who perceive themselves as generally having weak

behavioral coping stra
tegies. When discussing the influences on adoption of

major lifestyle change one
has to speak of "general" ability because many of the

specific challenges that
follow the major life style change are unanticipated at the

time of the decision.
Successful voluntary lifestyle change is often accompanied

by multiple changes
in how the individual interacts with her/his social and

physical environment. A
successful change in diet, for instance, will typically be

accompanied by changes
in shopping habits, changes in cooking habits, and changes

in where, when and
with whom to dine out. For predicting the success of major

lifestyle change ef
forts, one's ability to cope intrapsychically with

specific challenges seems less
relevant than one's general ability to have a behavioral

impact on one's environ
ment. This focus on "general" ability might seem

inappropriate for a discussion
of self-efficacy as it is classically defined (Bandura,

1977). This focus is consis
tent, however, with recent demonstrations that global

judgments about a subject's
ability to train employees or to influence organizational
performance can influ
ence their perception of their own organizational ability

(Bandura & Wood,
1989; Wood & Bandura, 1989) and subsequent organizational

performance. Description of Proposed Study For this
study, we followed the strategy of Ryckman, Thornton, and
Cantrell
(1982). A comprehensive range of 24 measures of personal

effectiveness was
administered to a community sample of young adults being

followed in a longi
tudinal study of growth and development. These measures

were submitted to a
hierarchical, confirmatory factor analysis in a random

half of the sample and
cross-validated in the other half. Based on the literature

discussed above, we hy
pothesized finding two second order factors of perceived

behavioral ability
(Perceived Behavioral Coping Ability) and perceived

cognitive control ability to
cope with environmental challenges (Perceived Cognitive

Control Ability).
Multiple assessments of personal effectiveness included

measures of perceived
ability to have a social impact, general assertiveness,

dating competence, social
support, depression, perceived loss of control, purpose in

life, and leadership
style. These were selected to provide a comprehensive range

of measures of the
subjects' cognitive and behavioral skills and emotional

states related to interper
sonal relations. We expected that Perceived Behavioral
Coping Ability would be
reflected in first-order factors of social impact efficacy

(perceived ability to have
a social impact), general assertiveness, social resources,

dating competence, and
leadership style because these represent ways of

behaviorally operating on the
environment. On the other hand, we expected that Cognitive

Control Coping
Ability would be reflected in constructs of social impact

efficacy, depression,
purpose in life, and perceived loss of control because

these involve internal cop
ing or cognitive qualities. Expected social impact efficacy

was expected to load
on both second-order factors because this measure included

cogmtIve and
behavioral coping items (Blatt, Quinlan, Chevron, McDonald,

& Zuroff, 1982).
Table 1
Description of Sample Characteristic N % Sex Male 221

30 Female 518 70 Age Mean 21.93 Range 19 24
Ethnicity Black 111 15 Hispanic 72 10 White 490 66
Asian 66 9 High School Graduate Yes 684 93 No 55 7
Living Situation Alone 28 4 Parents 343 46 Spouse 77
10 Spouse and Child 56 8 Cohabitation 67 9 Dormitory 40
5 Roommates 96 13 Other 32 4 Number of Children
None 619 84 One 106 14 Two or more 14 2 Current
Life Activity Military 23 3 Junior College 87 12
Four-year college 153 21 Part-time job 102 14 Full-time
job 343 46 Other 31 4 Income for Past Year None 71
10 Under $5,000 242 33 $5,001 to $15,000 334 45 Over
$15,001 92 12 The analyses proceeded in steps, first
testing for sex differences on the
measures. Then the adequacy of the hypothesized latent

measurement model was
tested followed by testing the second-order factor model
of the primary-order
latent variables. These analyses were conducted in the

derivation sample, and
then the second-order factor results were confirmed in the

cross-validation
sample, and in separate samples of men and women. METHOD

SUbjects Participants in this study were 739 young adults
who completed an eight-year
(fifth wave data assessment point) longitudinal study of

adolescent and young
adult development. Data were collected initially from 1,634

students in the
seventh, eighth, and ninth grades at II randomly-selected

Los Angeles County
schools. At this young adult follow-up, each participant

was paid $12.50 to
complete the questionnaire and all subjects were apprised

of a grant of confiden
tiality given by the U.S. Department of Justice.

Forty-five percent of the original
sample participated as young adults. The loss of subjects

due to attrition over
the eight years has been shown not likely to bias the
results adversely (Newcomb,
1986; Newcomb & Bentler, 1988a). Table I presents a

description of the sample. As evident, 30% were men and
70% were women, which parallels the sex distribution in the

initial sample and
does not reflect differential attrition by sex. Most were

employed full-time and
represented varied ethnic backgrounds. Additional

information is provided about
their current living arrangements, income, and current life

pursuits. When these
sample characteristics are compared with national samples

of young adults (e.g.,
Bachman, O'Malley, & Johnston, 1984; Miller et aI., 1983)

or other studies of
young adult populations (e.g., Donovan, Jessor, & Jessor,

1983; Kandel, 1984)
very similar patterns emerged. Consequently, we consider

this sample to be
reasonably representative of young adults in general.

Measures Table 2 presents a listing of the 24 variables
used in this study. They are
organized according to the latent construct they are

hypothesized to reflect. For
instance, the latent construct of Social Impact Efficacy is

assumed to generate the
variation in three observed indicators called inner

resources, independence, and
others' respect. For factors which tend to be

unidimensional in nature, three
measured-variable indicators were constructed from the

items to reflect the latent
factor or construct. This was done, since, as a rule, it

is recommended to have at
least three, highly correlated indicators to identify a

latent construct (e.g., Bentler
& Newcomb, 1986). This was done on the self-efficacy,

dating competence,
general assertiveness, and purpose of life factors.

Standard univariate statistics
for each variable are also given in the table. Below we

describe how each
variable was assessed in regard to the latent construct it

represents. Social impact efficacy. Three scales are used
to reflect the Social Impact
Efficacy construct. These were derived from the five-item

scale of efficacy
developed by Blatt et al. (1982). Responses to these five

items were given on a
five-point anchored rating scale that ranged from strongly

disagree (1) to
strongly agree (5). The items were factor analyzed and

found to reflect a unitary
construct (only one eigenvalue greater than 1.00 and all

factor loadings were
greater than .4 on the first unrotated factor). As a

result, these five items were
combined into three scales based on content. Inner

resources was assessed with a
single item-"I have many inner resources." Independence

was the average of
two items-"I am a very independent person" and "I set my

personal goals as
high as possible." Others' respect was the average of two

items-"Others have
high expectations of me" and "What 1 do and say has a

great impact on those
around me." General assertiveness and dating competence.

Dating Competence was as
sessed by three scales (dating I, dating 2, and dating 3)

derived from a nine-item
scale of social competence in dating situations. General

Assertiveness was asses
sed by three scales (assertive I, assertive 2, and

assertive 3) obtained from a nine
item social assertiveness scale. The total Dating

Competence and Social Asser
tiveness scales were developed by Levenson and Gottman
(1978), and in several
studies had quite good discriminant validity in both normal

and clinic samples.
In the derivation study the entire Dating Competence scale

had an internal con
sistency reliability alpha of .92, while for the General

Assertiveness scale the
alpha was .85. Latent constructs derived from these scales

have also been used in
a study of sexual behavior and responsiveness (Newcomb,

1984). Social resources. Three questions were asked to
determine the quantity or
amount of social supports as perceived in three life

contexts. The first item asked
"How many clubs, groups, or organizations do you belong to

(including church
groups)?" The second item asked "How many friends do you

reaJJy feel close
to?" And the third item asked "How many family members or
relatives can you
talk to about things personal to you?" Responses were given

on a rating scale
that ranged from none to nine or more. These items were

specifically developed
for this research project, but are similar to standard

measures of social support
that focus on amount of social resources and correlate

quite highly with satisfac
tion with social support from various types of social

networks (Newcomb &
Bentler, 1988b). Depression. The 20-item depression scale

from the Center for Epidemio
logic Study of Depression (CES-D) was completed by all

subjects. The devel
opment, validities, and reliabilities of the measure have

been reported elsewhere
(Husaini, Neff, Harrington, Hughes, & Stone, 1980; Radloff,

1977; Weissman,
Sholomskas, Pottenger, Prusoff, & Locke, 1977).

Participants were asked to rate
their frequency of occurrence for each of the 20 symptom

items during the past
week on a scale from none (0) to 5 -7 days (3). The 20

items were factor
analyzed in this sample and found to contain four distinct

factors, which is con
sistent with previous attempts to determine the factor

structure of the CES-D
(e.g., Clark, Aneshensel, Frerichs, & Morgan, 1981;

Radloff, 1977; Roberts,
1980). The four factors included positive affect, negative

affect, impaired moti
vation, and impaired relationships. Items were averaged

into the respective four
factors and were used as indicators of a general latent

construct of Depression. Perceived loss of control. Three
single-item variables are hypothesized to
reflect the construct of Perceived Loss of Control.

Subjects were asked to rate
their degree of agreement with three statements: (I) "I

feel I am not in control of
my life," (2) "} feel that whether or not I am successful

is just a matter of luck
and chance, rather than my own doing," and (3) "} feel
that others are running my
life for me." Responses were given on a seven-point

anchored rating scale that
ranged from strongly disagree (I) to strongly agree (7).
Cronbach's alpha for
these three items was .65. This construct assesses a

general lack of control over
life events, and has been validated in other samples and

studies (Newcomb, 1986;
Newcomb & Harlow, 1986). Purpose in life. The Purpose in

Life test (Crumbaugh, 1968; Crumbaugh &
Maholic, 1964, 1969) consists of 20 items designed to

assess one's level of or
purpose in life. Each item was rated on a seven-point

anchored rating scale rang
ing from strongly disagree (I) to strongly agree (7).

Previous research on the
Purpose in Life test indicated that it contains several

small primary factors and
one large general factor (Harlow, Newcomb, & Bentler,

1987). For purposes of
this study, the 20 items were randomly assigned into three

scales (PIL I, PIL 2,
and PIL 3) which were used as manifest indicators of a

latent construct of
Purpose in Life. Leadership style. Two personality scales,

ambition and leadership, were
used to reflect the construct of Leadership Style. These

traits were assessed
using a self-rating test modified for this research

program, but based on the
Bentler Psychological Inventory (BPI; Bentler & Newcomb,

1978; Huba &
Bentler, 1982). Although the BPI was developed with

multivariate methods, the
items have a high degree of face validity. Half of the

items for each trait are
reverse-scored to minimize response bias or acquiescence.

Four items were used
to assess each trait and each item was rated on a 5-point

bipolar scale. Thus,
each scale had a range of 4 to 20. The BPI has proved

useful in studies of marital
success and failure (Bentler & Newcomb, 1978), criminal

behavior (Huba &
Bentler, 1983), and adolescent substance use (Huba &

Bentler, 1982). The
period-free test-retest reliability for ambition was .72

and the reliability for
leadership was .71 (Stein, Newcomb, & Bentler, 1986).

Analyses Our first set of analyses use point-biserial
correlations to test for mean differ
ences between men and women on each of the 24 variables.

We next use a con
firmatory factor analysis with latent variables to evaluate

the adequacy of the
hypothesized factor structure (e.g., Bentler, 1980; Bentler

& Newcomb, 1986), in
a random half of the sample, which we call the derivation

sample. An inspection
of the skew and kurtosis estimates for the 24 observed

measures indicates that
they are relatively normally distributed. As a result we

will use the maximum
likelihood structural model estimator, which requires

multivariately normal data
(e.g., Bentler, 1983, 1986). If the initial hypothesized

model does not adequately
reflect the data (which is common in models with many

variables and many sub
jects), we will modify the model until an acceptable fit is
achieved, in a manner
which will not disturb the critical features of the model.

These empirical model
modifications will be guided by the multivariate

Lagrangian Multiplier test for
adding parameters and the multivariate Wald test for

deleting parameters (Bentler
& Chou, 1986). Once this is accomplished, we will attempt

to confirm our two
hypothesized second-order constructs in this model, making

modifications where
necessary. This final model will be tested separately in

the other random half of
our sample (called the cross-validation sample), as well as

the samples of men
and women to determine whether the second-order factor

structure is an accurate
representation in these samples. The final model will also

be tested in the origi
nal derivation sample, but without the empirical

modifications to establish
whether the model modifications may have distorted or

biased the final results. RESULTS Sex Differences Mean
differences between men and women on the 24 variables were
tested
using point-biserial correlations. Males were coded I and

females were coded 2,
so that a positive correlation indicates that the women had

the larger value and a
negative correlation indicates that the men had the larger

value. These mean dif
ference correlations are presented in the right-hand column

of Table 2. Of the 24 variables, significant mean
differences were found on 14 of them.
These differences indicate that the women, compared to the

men, felt that they
had fewer inner resources, less independence, less respect

from others, less as
sertiveness (on all three scales), a smaller number of

friends they could rely on,
less positive affect, more negative affect, more impaired

motivation, others con
trolled her life more, slightly less dating competence

(only one scale significantly
different), less ambition, and fewer leadership qualities.

Although there were
many mean differences between the men and women, the

magnitude of these dif
ferences was quite small. For instance, the largest

difference accounted for only
four percent of the variance between groups (on ambition).

Based on these rather
small in magnitude mean differences between men and women.

and previous re
sults indicating that there were not different factor

structures for men and women
on social support and loneliness variables (Newcomb &

Bentler, 1986) and on
physical health status indicators (Newcomb & Bentler,

1987), we will collapse
across sex for the bulk of the remaining analyses.

However, we will test our final
model in the separate samples of men and women to

determine whether we may
have obscured any important findings by combining the men

with the women.
Table 2
Summary o/Variable Characteristics and Sex Mean Difference

Tests Sex
FactorN ariable Mean Range SD Skew Kurtosis Difference rpb

a
Social Impact Efficacy
Inner resources 3.87 1-5 .75 -.55 .61 -.18***
Independence 3.86 1-5 .76 -.76 .41 -.11 **
Others respect 3.64 1-5 .62 -.39 .56 -.08*
General Assertiveness
Assertive 1 8.89 4-13 1.84 -.13 -.38 -.13**
Assertive 2 9.07 4-14 1.90 .03 -.32 -.18***
Assertive 3 8.74 4-14 1.87 .05 -.24 -.09*
Social Resources
Number of family members 3.48 0-9 2.44 .91 .12 .02
Number of friends 4.04 0-9 2.34 .71 -.10 -.14***
Number of organizations 1.34 0-9 1.36 1.21 2.54 -.06
Depression (CES-D)
Positive affect 2.34 0-3 .64 -.96 .32 -.10**
Negative affect .63 0-3 .64 1.20 1.20 .15***
Impaired motivation .71 0-2.75 .46 .79 .79 .08*
Impaired relationships .34 0-3 .47 1.72 3.30 .00
Purpose in Life
PIL I 5.57 2-7 1.80 -.81 .22 -.04
PIL2 5.30 2.4-7 1.80 -.58 .06 -.06
PIL 3 5.41 2-7 1.77 -.57 .13 -.05

Perceived Loss of Control
Not in control 2.26 1-7 1.51 1.31 1.05 -.02
Powerless 2.28 1-7 1.45 1.20 .82 .03
Others control life 2.73 1-7 1.50 .87 .01 .11 **
Dating Competence
Dating 1 8.84 4-13 1.75 -.06 -.28 -.08*
Dating 2 9.25 3-13 1.82 -.19 -.23 -.03
Dating 3 9.96 4-14 1.87 -.14 -.17 -.05
Leadership Style
Ambition 14.52 4-20 3.53 -.41 -.49 -.20***
Leadership 14.29 5-20 2.83 -.13 -.30 -.16***
Note. a Males were coded I and females were coded 2, so

that a positive point-biserial correlation indicates that
the females had the larger value. *p < .05; **p < .01;
***p < .001. First-Order Latent Factor Model This first
sequence of models are tested on a random half (n = 370) of
our
total sample, which we call our derivation sample. In the

initial confirmatory fac
tor model, the eight latent constructs were hypothesized to

"cause" or generate
the variation in the 24 observed variables. The factor

structure of this first model
was "pure" in that each observed variable was allowed to

load on only one latent
construct. For instance, inner resources was assumed to be

an indicator only of
Social Impact Efficacy. This assumption of mutual

exclusivity may be an overly
constrained imposition on the model, since many of the

variables are concep
tually similar and may in fact reflect more than one
underlying quality. To identify the model all factor
loadings were freed, the variances of the con
structs were fixed at unity, and all factor

intercorrelations were allowed to be
freely estimated. This initial confirmatory factor model

did not fit the data to an
acceptable degree, chi2 (df = 224, N = 370) = 480.17, p <

.001, NR (normed fit
index: Bentler & Bonett, 1980) = .86. Latent-factor

intercorrelations for this mo
del are presented in the upper triangle of Table 3. All

hypothesized factor load
ings were highly significant, p < .00]. The NFl was

sufficiently large to suggest
that an acceptable model could be achieved by adding

several small empirically
determined parameters that were not hypothesized in the

initial model. Based on an examination of selected
modification indices for additional
factor loadings and correlated uniquenesses (Bentler &

Chou, 1986), five non-hy
pothesized factor loadings and 22 correlations among

manifest variable residuals
were added to the model. With these additions the mode]

adequately fit the data,
chi 2 (df = 196, N = 370) = 181.97, p = .76, NFl = .95.

This new model was a sig
nificant improvement over the initial model (p < .(01). A

summary of all model
fit statistics and difference chi2 tests are given in Table

4. To test whether the ad
dition of these empirically-determined parameters

distorted the substantive inter
pretation of the model, the latent-factor
intercorrelations from the initial model
were correlated with those obtained in the final, modified

model. These param
eters were correlated greater than .99. As a result, the

final model was not
considered biased due to the model modifications.

Standardized factor loadings and residual variances for the
final first-order
confirmatory factor model are given in Figure I. The

rectangles represent the ob
served variables, the large circles indicate the latent

constructs, and the small cir
cles reflect residual variances of the observed variables.

These five non-hypothe
sized factor loadings tend to be small in magnitude (only

one is over .40) and all
are in interpretable directions. For instance, general

assertiveness also negatively
influenced the perception that others control your life.

The latent-factor intercor
relations for this model are given in the lower triangle

of Table 3. The absolute
value of the correlations ranged from a low of .06 to a

high of .88, and II were
higher than .5. In other words, many of the constructs

appear to be highly
correlated and may reflect higher-order factors, as

originally hypothesized. F i g u r e 1 F i n a l c o n f i
r m a t o r y f a c t o r a n a l y s i s m o d e l f o r t
h e d e r i v a t i o n s a m p l e . L a r g e c i r c l e
s r e p r e s e n t l a t e n t f a c t o r s , r e c t a n
g l e s o b s e r v e d v a r i a b l e s , a n d s m a l l
c i r c l e s r e s i d u a l s . N o t d e p i c t e d i n
t h e f i g u r e f o r r e a s o n s o f c l a r i t y a r
e t w o h e a d e d a r r o w s ( c o r r e l a t i o n s )
j o i n i n g a l l p o s s i b l e p a i r s o f l a t e n
t f a c t o r s . P a r a m e t e r e s t i m a t e s a r e
s t a n d a r d i z e d a n d r e s i d u a l v a r i a b l
e s a r e v a r i a n c e s . S i g n i f i c a n c e l e v
e l s w e r e d e t e r m i n e d b y c r i t i c a l r a t
i o s ( U p < . 0 1 ; * * * p < . 0 0 1 ) .
.64 Inner resources
.64 Independence .60 .60
.66 Others respect .74
.38 Assertiv e 1 .79
.45 Assertiv e 2 .74 .59 Assertiv e 3 .66 Number of familv

members
.67 Numbe r of friends
,67
.92 Number of organizations
.47 Positive affect
.22 Negative affect .57 .57 .-.40 .88 Impaired motivation
.42' Impaired relationships .59, .76 .48 .2 6
.37 PIL1
.39 PIL2
.31. PIL 3
.50 Not in control .78 .8 3 .85
.74 Powerless
.58 Others control ffe .51 7 1
.21 Dating 1 Dating 2.37, .89 .79 ,63 .73 Leadership

Ambition Dating 3 .48
.59,
.44 Social Impact Efficacy Gemeral Assertiveness .09 Social

Resources Depression (CES-D ) .43 Purpose in Life -.35
Perceived Loss of Control .72 Dating Competence Leadership
Style -.28 .57 .51
50
Table 3
Factor Intercorrelations Between the Initial (Upper

Triangle) and Final (Lower
Triangle) First-Order Confirmatory Factor Models
Factor II III IV V VI VII VIII Social Impact Efficacy

1.00 .65 .44 -.44 .71 -.59 .51 .62
II General Assertiveness .60 1.00 .30 -.31 .50 -040

.58 .67
III Social Resources .40 .31 1.00 -.37 .50 -.30 048 .20
IV Depression (CES-D) -.44 -.34 -.29 1.00 -.59 .73 -.35

-.22
V Purpose in Life .78 .46 .50 -.53 1.00 -.88 .54 .30
VI Perceived Loss of Control -.61 -.39 -.25 .61 -.84

1.00 -.48 -.32
VII Dating Competence .56 .61 049 -.32 .52 -A 1.00 .48
VIII Leadership Style .61 .56 .12 -.06 .27 -.25 AI

1.00
Note. r between initial and final correlations> .99. All

correlations are significant at p < .01. Second-Order
Confirmatory Factor Models Based upon our theoretical
position, which hypothesized that two second
order factors should underlie the construct of

self-efficacy, two second-order
factors were introduced into the confirmatory factor

analysis. One second-order
factor represented Perceived Behavioral Coping Ability and

was reflected in
loadings allowed on social impact efficacy, general

assertiveness, social
resources, dating competence, and leadership style. The

other second-order
factor reflected Perceived Cognitive Control Coping Ability
with hypothesized
factor indicators of social impact efficacy (the only

first-order construct to load
on both second-order factors), a lack of depression,

purpose in life. and a lack of
perceived loss of control. A model was tested which

included the two second-order factors as defined
above. Perceived Behavioral Coping Ability and Perceived

Cognitive Control
Coping Ability were allowed to correlate freely. Three

additional empirically
determined correlations were included between pairs of

first-order factor resid
uals: Depression and leadership style, social impact

efficacy and leadership
style, and depression and perceived loss of control. The

factor residual of
purpose in life was fixed at zero in order to prevent it

from being estimated as
negative. This model adequately fit the data and was not
significantly different from
the first-order confirmatory factor models, even though 16

fewer parameters
were necessary to represent the first-order latent factor

intercorrelations (see
summary of fit indices in Table 4). This model is

graphically depicted in I
Figure 2 omitting the observed variables for clarity.

Parameter estimates are
standardized and residual variables are variances.

Table 4
Summary of Fit Statistics
Model
1. Initial CFAa 480.17
2. Final CFAb 191. 97
Model 1-2 difference 288.20
3. Two second-order factors on Model 2 211.13
4. Two second-order factors on Modell 501.04
5. Second-order factors correlated at unity 278.89
6. Separate cross-validation sample-Model 3 358.54
7. Males only, Model 3 219.29
8. Females only, Model 3 352.51 Degrees of p Value

Freedom Derivation Sample (N = 370) 224 <.001 196 .76
28 <.001 212 .50 16 .26 240 <.001 16 .18 213 .001 1
<.001 Cross-Validation Samples 212 <.001 212 .35 212
<.001 Normed Fit Index .86 .95 .94 .87 .92 .89 .90
.92
Note. a CFA = Confirmatory factor analysis. b Modified by

adding 22 correlated residuals and 5 nonhypothesized
factor loadings. Perceived Behavioral Coping Ability and
Perceived Cognitive Control
Coping Ability were correlated .66, indicating a moderate

association between
them (44% common variance), while retaining their own

uniqueness. This cor
relation resembled the correlation of .44 that Folkman and

Lazarus (1980)
observed between emotion-focused and problem-focused coping

and the range of
correlations (.48-.37) between perceived coping and

cognitive control efficacy
that Ozer and Bandura (1990) recently reported. (These

studies measured var
iables rather than latent constructs, which might account

for the smaller size).
The largest additional correlation was between the

residuals of Social Impact chi2
Efficacy and Leadership Style, indicating that the two

second-order factors did
not account for the entire association between these two

constructs. This
association may reflect an additional second-order factor

for these two constructs,
over-and-above the relationship accounted for by the

Perceived Behavioral
Coping Ability factor. A similar possibility may exist for

Depression and
Perceived Loss of Control. These were not tested because

two-indicator factors
tend to be very unstable, and the fit and interpretability

of the model seem to be
quite good as it stands. In order to determine whether the

two second-order factors of self-efficacy
that we have identified are in fact separate constructs,

an additional more restrict
ed model was tested. In this model the correlation between

the two second-order
factors was fixed at 1.0, operationalizing the hypothesis

that they are assessing
the same quality. This model did not accurately reflect the
data and was signifi
cantly worse when compared to the previous model which
allowed the two
second-order constructs to be unique (see summary of fit

statistics in Table 4). Finally, we tested the
second-order factor model in the initial confirmatory
model that did not include the additional five factor

loadings nor the 22 corre
lated uniquenesses. All significant relationships were

retained and the resultant
model was not significantly different from the initial

model. Thus, we conclude
that the model depicted in Figure 2 is an accurate

portrayal of the data that is not
biased or distorted due to model modifications.

Cross-Validation of the Second-Order Factor Model This
final model was tested in the separate, untouched random
half of the
total sample, as well as the separate samples of men and

women. Fit indices for
these runs are given in Table 4 (Models 6, 7, and 8). In

each of these three sam
ples, all hypothesized factor loadings and second-order

factor results were
significant. Although only the male sample fit the model

according to the p-value crite
rion, all three models had NFIs greater than .89,

indicating that each fit the data
reasonably well. Similarly. the ratio of chi 2 to degrees

of freedom was
consistently under 2.0, also reflecting an excellent degree

of fit. All hypothesized
factor loadings on the first-order factors were significant

in each of the three
sample partitions. Table 5 presents the standardized

parameter estimates for the
second-order factors for the derivation sample, as well as
the cross-validation,
female, and male samples. Although the magnitude of the

factor loadings varied
somewhat, the general patterns of association were

remarkably similar. The cor
relation between the two second-order factors was

consistently in the .64 to .66
range. Using the cross-validation method suggested by

Cudeck and Browne
(1983), the specific model developed in the derivation

sample was used in the
cross-validation sample (by imposing identical

parameterization and parameter
estimates) and accounted for 86% of the variance of their

new data. Although the
fit was worse in this second sample, the decrement was not
substantial, and more
important, the substantive conclusions (i.e.,

interpretation of parameter estimates)
was virtually identical in the derivation and

cross-validation sample, when
parameters were estimated freely (Table 5). As a result, we

conclude that the
factor model presented in Figure 2 is equally

representative of men and women,
and thus does not differ by sex of the subject, and has
been cross-validated in a
separate sample.
Table 5
Summary of Second-Order Factor Parameters for Several

Sample Partitions
Parameter Derivation Cross-Validation Female Male Sample
Sample Sample Sample Second-Order Factor Loadings
Behavioral Coping Efficacy
Social impact efficacy AO .28 .32 .31
General assertiveness .73 .74 .74 .66
Social resources .57 A9 .60 A7
Dating competence .81 .64 .69 .81
Leadership style .50 .50 .56 A6
Cognitive/Emotional Coping
Efficacy
Social impact efficacy .51 .59 .55 .59
Depression -.53 -A7 -A7 -.56
Purpose in life 1.00 .97 1.00 1.00
Perceived loss of control -.85 -.93 -.91 -.85 Factor

Correlations
Perceived Behavioral Coping .66 .64 .66 .66
Ability with Perceived
CognitivelEmotional Ability
Depression (R)a with perceived
loss of control (R) AI .25 .35 .32
Social impact efficacy, (R)
with leadership style (R) A8 A9 .52 .31
Depression (R) with
leadership style (R) -.20 -.12 -.15 -.17
Note. a (R) denotes factor residual.
Figure 2 Final second-order factor model for the derivation

sample. The large circles are latent constructs (the two
on the right-hand side are second-order factors); the
small circles represent factor residuals. Two-headed arrows
are correlations. Parameter estimates are standardized and
residual variables are variances. Significance levels were
determined by critical ratios (**p < .0 I; ***p < .00 I).
.30 Socia l Impact SelfEfficacy .47 Genera l Assertiveness
.,40 .73 Behavioral Coping Ability .57. .6 7 Social
Resources Depression (CES-D) .72 -.5 3 .66 .51 Cognitive
Control Ability 1.00 .41.20.48 Purpose in Life Perceived
Loss of Control .28 .35 .81 Dating Competence Leadership
Style .50 .73 .28 .2 8 DISCUSSION
Affirmation of the Distinction Between Perceived Ability to
Self-Regulate Cognitively and Perceived Ability to Cope

Behaviorally
With Environmental Challenges Our results confirm the

validity of distinguishing between beliefs about one's
ability to regulate cognitions in response to challenges

associated with accomp
lishing desired goals and beliefs about one's ability to

have an impact on the
environment to accomplish desired goals. Those constructs

that are primarily
intrapsychic, such as depression, purpose in life, and

perceived loss of control
loaded heavily on the Perceived Cognitive Control Ability

factor, but did not load
on the Perceived Behavioral Coping Ability factor. Those

constructs that con
cern active involvement with one's social environment, such

as general asser
tiveness, leadership, and dating competence, on the other

hand, loaded heavily on
the Perceived Behavioral Coping Ability factor, but not on

the Perceived
Cognitive Control factor. The empirical results obtained

in this study confirm the usefulness of a theo
retical distinction that is already current and frequently
applied in the literature
on coping with stress. This report, of course, goes beyond

the literature on coping
to justify elaborating the self-efficacy construct to

include the distinction between
perceived cognitive control and perceived behavioral

abilities. The findings from
a recent prospective study involving these two

distinguishable percepts of ability
(Ozer & Bandura, 1990) further affirm the importance of

distinguishing between
cognitive control and behavioral coping ability. Ozer and

Bandura noted in this
study of women mastering self-defense skills that perceived

[behavioral) coping
self-efficacy developed more rapidly during the period of

training than did cogni
tive control self-efficacy. One explanation for this

finding is that the behavioral
and attitudinal changes engendered by 22 contact hours of

training in thwarting
simulated physical assaults were too many, too disparate,

and too emotionally
disturbing to be fully apprehended and integrated at the

time of skill acquisition.
The resulting dissociation of perceived cognitive control

and perceived
[behavioral) coping self-efficacy during the skill

acquisition phase was tempo
rary. By the 6-month follow-up the positive correlation

between cognitive
control efficacy and [behavioral) coping self-efficacy was

restored (r = .37),
despite sustained increases in self-defense efficacy and
sustained decreases in
perceived vulnerability and anxiety. It is as if time and

effort were required to
alter existing patterns of cognitive control following

sudden major changes in
perceived behavioral coping self-efficacy. If these

findings are generalizable to
other efforts at mastering complex behaviors, they suggest

the need to examine in
more detail the nature and sequencing of self-efficacy

beliefs prior to, during, and
following the acquisition of mastery.
Health Consequences of High Intrapsychic Efficacy
Relative to Problem-Focused Efficacy Individuals high in

Perceived Cognitive Control Coping Ability, relative to
Perceived Behavioral Coping Ability, may end up merely

palliating problems
that could be resolved behaviorally. For example, a woman

suffering from an
unhappy marriage may successfully reduce that unhappiness

via psychotherapy
that boosts her self-esteem, thereby increasing her ability

to avoid intrusive, self
denigrating thoughts, without doing anything about

increasing her perceived
ability to influence her husband's insensitivity to her

needs. Individuals who perceive that they have little
control over their environment
can nevertheless increase their probability of achieving a

healthful lifestyle
change by maximizing their Perceived Cognitive Control. An

example would be
an individual working in a tobacco processing plant who
would like to quit
smoking. Such an individual would find few opportunities

for changing the
social and physical environment to make it more conducive

to quitting smoking,
but might nevertheless embark on a plan to become an

exsmoker by relying on
primarily cognitive control strategies for coping with

temptations to smoke.
Cognitive strategies such as thought management,

self-reward, and distractive
thoughts for reducing urges to smoke could be employed to

accomplish a suc
cessful transition to being a nonsmoker even though the

individual's Perceived
Behavioral Coping Ability was low. Conversely, smokers who

feel that they have little "willpower" and who are
hostile to conventional intrapsychic approaches to behavior

change can never
theless be successfully counseled to effect a successful

change to being a non
smoker by being encouraged to adopt behaviors seriatim

that cumulatively are
antithetical to the smoking habit. A program of steadily

increasing physical
activity, for instance, is likely to improve one's chances

of becoming a long-term
nonsmoker because aerobic exercise is inherently

inconsistent or incompatible
with the cigarette smoking lifestyle (e.g., Koplan, Powell,

Sikes, Shirley, &
Campbell, 1982). Successful adoption of behaviors that are

inconsistent with
smoking should lead to increased access to images of the

self as a nonsmoker,
with an attendant increase in perceived ability to abstain

from smoking (Kazdin,
1979).
Implications of the Distinction/or Interventions Designed

to Change
Perceptions 0/ Self-Efficacy Interventions designed to

influence individuals' self-efficacy with respect to
a desirable health-related lifestyle change have not

distinguished between
Perceived Behavioral Coping Ability and Perceived Cognitive

Control Ability.
In the few instances where distinctions have been made

between the use of cogni
tive or behavioral coping strategies, there have been

notable differences in the
contribution of behavioral and cognitive coping to

explaining the behavior
change. The literature on the behavior change involved in

drug use cessation, to
take one example, includes reports of contrasts between

behavioral and cognitive
methods of coping with temptations to return to drug use.

One such applied
study found that recidivists among would-be exsmokers

reported relying more
heavily on behavioral coping than did the more successful

exsmokers who con
tinued to abstain from smoking (Shiffman, Reed, Maltese,

Rapkin, & Jarvik,
1985). Another study investigated the determinants of

cessation of heroin use
and found a similar advantage for cognitive coping relative

to behavioral coping
strategies (Chaney & Roszell, 1985). The findings

reported above could be an artifact of the intervention
model
used, namely the Relapse Prevention model (Marlatt &

Gordon, 1985). Much of
the work on relapse prevention has focused on how to equip

individuals with
self-efficacy percepts that would help them cope in

situations that pose a high
risk of recidivism. The focus however, has not been on

increasing the individ
ual's perceived ability to reduce exposure to high-risk

situations, but rather on
the individual's perceived ability to reduce the experience

of stress in high-risk
situations. In other words, the focus has been such as to

exaggerate the
importance of perceived cognitive control for coping with

high-risk situations
relative to the importance of perceived ability to respond

behaviorally for
avoiding or escaping from high-risk situations. The

superiority of enhancing percepts of Behavioral Coping
Ability rather
than enhancing percepts of Cognitive Control Coping

Ability in lifestyle change
programs is suggested by some multi-year follow-ups of

heroin addicts. In their
12-year follow-up of the effects of treatment of 405 black

and white male opiate
addicts, Simpson, Joe, Lehman, and Sells (1986) concluded

that the most predic
tive determinants of long-term continued abstinence were

primarily behavioral:
Avoiding old drug-using friends and old hangouts,

developing new friendships
with nonusers, and establishing new family ties and new

work habits. In their
review of the determinants of spontaneous remission from

substance use, Stall
and Biernacki (1986) arrived at similar conclusions. These

results suggest that
would-be ex-addicts with strong beliefs about their ability

to cope behaviorally
will experience higher rates of long-term abstinence than

would-be ex-addicts
who may have strong beliefs in their ability to cope

cognitively but weak beliefs
about their ability to cope behaviorally. Percepts of

ability have been shown to be important determinants of
effort
and achievement (Bandura, 1986). Failure to distinguish

between perceptions of
cognitive control ability and perceptions of behavioral

coping ability, however,
could mask important information about the processes by

which actions, beliefs
and perceptions of ability influence each other. Two

examples are given below
where potentially important applications of social

cognitive theory may be limit
ed by the failure to make this distinction.
Conjecture Relating to Adolescent Maturation and Risk of

Drug Abuse An important, unexplained phenomenon in the
literature on drug abuse onset
is the "window of vulnerability," namely, the relatively
few teenage and young
adult years when individuals are at risk of adopting a drug

abusing lifestyle
(Abelson, Fishburne, & Cis in, 1980; Johnston, O'Malley, &

Eveland, 1978;
Kandel & Logan, 1984). Lifestyle drug abuse rarely begins

earnestly before
adolescence and almost never manifests de novo after age

25. The "Just Say No"
drug prevention program is premised on the belief that

success in dissuading
teenagers from starting drug abuse during the teenage years

will prevent drug
abuse at any age. Why, over a lifespan of 72-78 years,

should the average American only be at
risk of lifestyle drug abuse between the ages of 13 and

25, with peak onset during
the high school years? Part of the answer may be

facilitated by distinguishing
between cognitive control and behavioral coping ability.

The behavioral coping
ability of children is generally limited to secondary

control (e.g., Rothbaum &
Weisz, 1989) because of their societally-mandated

dependency on their parents
and because of their lack of life skills. Children's

maturation is marked more by
increases in their cognitive control ability (e.g.,

distractive thoughts) than in their
behavioral coping skills (e.g., progressive goal-setting to

achieve mastery over
challenge; Altshuler & Ruble, 1989). The transition from

childhood to adulthood is almost inevitably accompanied
by increases in behavioral coping ability. The life skills

that are acquired include
decision-making skills, communication skills, dating

skills, and employment
skills. At the beginning of the transition, these

behavioral skills are uniformly
absent but young adolescents become increasingly aware of

the need to acquire
them (Katz & Zigler, 1967). There is considerable distress

and anxiety that
accompanies adolescents' increasing realization of the

need for life skills in the
immediate absence of their acquisition. This distress and

anxiety are palliated in
teenagers performing well in school by

societally-administered reassurances that
their career trajectory is favorable and that, by

implication, the teenagers need not
fear a characterological inability to acquire the

necessary life skills. For these
success-bound teenagers, positive self-statements are

easily accessible as anti
dotes to the inevitable anxiety that their immaturity

occasions. For many
teenagers not performing well in school and otherwise not

receiving societal reas
surances concerning future expectations of success,

however, only the actual
acquisition of life skills will permanently reduce the

fear that they will never be
fully accepted as autonomous, responsible adults. The

literature, shows, in fact,
that drug abuse-prone teenagers are characterized by a
syndrome of "accelerated
maturity," (Gritz, 1977), which manifests in precocious

sexual behavior, mar
riage, cessation of schooling, and employment. Despite the

uniform absence of
life skills at the beginning of adolescence, only a

minority go on to adopt a life
style habit of drug abuse. The at-risk teenagers who

successfully avoid drug
abuse are those who can through cognitive control alone

reduce their immaturity
associated anxiety to acceptable levels. At-risk teenagers

who successfully avoid
drug abuse tend to come from intact families, suggesting

that family social sup
port can strengthen self-percepts of ability to control

immaturity-associated anxi
ety. For at-risk teenagers without the requisite cognitive

control skills, their
immaturity-associated anxiety is functionally (but only

intennittently) palliated
by regular administration of psychoactive drugs,

especially nicotine and alcohol.
Kaplan, Martin, and Robbins (1982) demonstrated

prospectively that non-drug
abusing teenagers with low self-esteem were significantly

more likely in future
years to become drug abusers than their non-drug abusing

agemates with high
self-esteem. Orug abuse, therefore, can be viewed as a

functional way to medi
cate for intrapsychic discomfort. By the time of onset of

adulthood, most
individuals have demonstrated successful mastery of at
least the rudiments of the
most important life skills. With the ebbing of

immaturity-associated anxiety,
there is decreased need for psychoactive agents to provide

functional relief. By
the time of young adulthood, unfortunately, many

individuals have become phy
siologically dependent on their chosen drugs and cannot,

therefore, stop using the
drug just because the original need for the drug has
disappeared.
A Conjecture Concerning Self-Efficacy Gender Differences
in Sex-Role Socialization The distinction between

perceived cognitive control ability and perceived
behavioral coping ability may similarly shed light on the

origins of observed dif
ferences between men and women in mastery of a variety of

life's challenges,
including occupational achievement (Austin & Hanisch, 1990)

and weight con
trol (Jeffery, French, & Schmid, 1990). Women's continuing

preference for
teaching, nursing, and childcare, and men's continuing

preference for construc
tion, community safety (policy, fire, paramedics), and

surgery, are consistent
with women relying more on cognitive control coping and men

relying more on
behavioral coping. Similarly, women's relative reluctance

to adopt increased
physical activity and their preference for relying on

willpower relative to men as
a strategy for maintaining desirable weight is also
consistent with women relying
on cognitive control and men relying on behavioral coping.

This difference
between men and women may help to explain why women to view
weight as less
controllable than men (Jeffery et aI., 1990), given the

clear long-tenn advantage
that exercise represents as a weight loss strategy (King,

Frey-Hewitt, Oreon, &
Wood, 1989; Koplan et aI., 1982). In her review of the

literature on sex role socialization, Weitz (1977) cited
studies of adult communication patterns in same-sex groups

in which it was
observed that women tended to be socioemotional whereas

men tended to be
task-oriented. Although similar communication patterns were

not observed in
children, Weitz noted a consistent association of activity

and aggression in boys
and not in girls. Her evidence suggested that boys were

encouraged to combat
their frustrations behaviorally and that girls were more

encouraged to palliate
their frustrations through cognitive control strategies.

Consistent gender dif
ferences in adult perfonnance could well have roots in the

different coping
strategies that boys and girls are encouraged to develop.

CONCLUSION Further research on how perceived cognitive
control ability and perceived
behavioral coping ability vary among individuals, among

situations, and within
individuals over time seems warranted. Investigating the
relative importance of
perceived cognitive control and perceived behavioral coping

ability in therapy
mediated lifestyle change and spontaneous, unaided

lifestyle change would seem
especially worthwhile. This distinction would also seem

useful in illuminating
more clearly why there exist differences in mastery between

men and women,
and in better understanding what contributes to the

youthful decision to adopt a
Abelson, H., Fishburne, P., & Cisin, I. (1980). The

national survey on drug abuse:
Main findings 1979. Rockville, MD: National Institute on

Drug Abuse.
Altshuler, 1. L., & Ruble, D. N. (1989). Developmental

changes in children's
awareness of strategies for coping with uncontrollable

stress. Child Development, 60, 1337-1349.
Austin, 1. T., & Hanisch, K. A. (1990). Occupational

attainment as a function of abili
ties and interets: A longitudinal analysis using project

TALENT data. Journal of
Applied Psychology, 75,77-86.
Bachman, 1. G., O'Malley, P. M., & Johnston, L. D. (1984).

Drug use among young
adults: The impacts of role status and social environment.

Journal of Personality
Bandura. A. (1977). Self-efficacy: Toward a unifying theory

of behavioral change.

action: A social cognitive
Bandura, A., & Wood, R. (1989). Effect of perceived

controllability and performance
standards on self-regulation of complex decision making.

Bentler, P. M. (1980). Multivariate analysis with latent

variables: Causal modeling.
Bentler, P. M. (1983). Some contributions to efficient

statistics in structural models:
Specification and estimation of moment structures.

Psychometrika, 48, 493-517.
Bentler, P. M. (1986). Structural modeling and

Psychometrika: An historical per
spective on growth and achievements. Psychometrika, 51,

35-51.
Bentler, P. M., & Bonett, D. G. (1980). Significance tests

and goodness of fit in the
analysis of covariance structures. Psychological Bulletin,

88, 588-606.
Bentler, P. M., & Chou, C. P. (1986, April). Statistics

for parameter expansion and
construction in structural models. Paper presented at the

American Educational Re
search Association meeting, San Francisco.
Bentler, P. M., & Newcomb, M. D. (1978). Longitudinal

study of marital success and
failure. Journal of Consulting and Clinical Psychology, 46,

1053-1070.
Bentler, P. M., & Newcomb, M. D. (1986). Personality,
sexual behavior, and drug use
revealed through latent variable methods. Clinical

Psychology Review, 6, 363-385.
Blatt, S. 1., Quinlan, D. M., Chevron, E. S., McDonald,

c., & Zuroff, D. (1982). De
pendency and self-criticism: Psychological dimensions of

depression. Journal of
Chaney, E. F., & Roszell, D. K. (1985). Coping in opiate

addicts maintained on metha
done. In S. Shiffman & T. A. Wills (Eds.), Coping and

substance use (pp. 267-291).
Orlando, FL: Academic Press.
Clark, V. A., Aneshensel, C. S., Frerichs, R. R., &

Morgan, T. M. (1981). Analysis of
effects of sex and age in response to items on the CES-D

scale. Psychiatry Research,
5,171-181.
Courtois, C. A. (1988). Healing the incest wound. New York:

Norton.
Crumbaugh, 1. C. (1968). Cross-validation of purpose in

life test based on Frankl's
concepts. Journal of Individual Psychology, 24, 74-81.
Crumbaugh, 1. C., & Maholick, L. T. (1964). An

experimental study in existentialism:
The psychometric approach to Frankl's concept of neogenic

neurosis. Journal of
Crumbaugh, J. C., & Maholick, L. T. (1969). Manual of

instructions for the purpose in
life test. Munster, IN: Psychometric Affiliates.
Cudeck, R., & Browne, M. W. (1983). Cross-validation of

covariance structures.
Multivariate Behavioral Research, /8, 147-167.
Donovan, J. E., Jessor, R., & Jessor, L. (1983). Problem

drinking in adolescence and
young adulthood: A follow-up study. Journal of Studies on

Alcohol, 44,109-137.
Folkman, S., & Lazarus, R. S. (1980). An analysis of coping

in a middle-aged com
munity sample. Journal of Health and Social Behavior,

21,219-239.
Gritz, E. R. (1977). Smoking: The prevention of onset. In

Jarvik, M. E. et al.
Research on smoking behavior (pp. 290-307). NIDA Research

Monograph #17,
DHEW Pub. No. (ADM) 78-581.
Harlow, L. L., Newcomb, M. D., & Bentler, P. M. (1987).

Purpose in life test as
sessment using latent variable methods. British Journal of

Clinical Psychology, 26,
235-236.
Huba, G. J.. & Bentler, P. M. (1982). A developmental

theory of drug use: Derivation
and assessment of a causal modeling approach. In B. P.

Baltes, & O. G. Brim, Jr.
(Eds.), Life-span development and behavior, Volume 4 (pp.

147-203). New York:
Academic Press.
Huba, G. J., & Bentler, P. M. (1983). Causal models of the

development of law
abidance and its relationship to psychosocial factors and

drug use. In W. S. Laufer &
J. M. Day (Eds.), Personality theory, moral development,

and criminal behavior (pp. 165-215). Lexington, MA: Heath.
Husaini, B. A., Neff, J. A., Harrington, J. B., Hughes, M.

D., & Stone, R. H. (1980).
Depression in rural communities: Validating the CES-D

scale. JournalofCommun
ity Psychology, 7, 137-146.
Jeffery, R. W., French, S. A., & Schmid, T. 0. (1990).

Attributions for dietary failures:
Problems reported by participants in the Hypertension

Prevention Trial. Health Psy
chology, 9, 315-329.
Johnston, L., O'Malley, P., & Eveland, L. (1978). Drugs

and delinquency: A search
for causal connections. In D. B. Kandel (Ed.),

Longitudinal research on drug use:
Empirical findings and methodological issues (pp.

137-156). Washington, DC:
Hemisphere-Wiley.
Kandel, D. B. (1984). Marijuana users in young adulthood.

Archives of General
Psychiatry, 4/, 200-209.
Kandel, D. B., & Logan, J. A. (1984). Patterns of drug use

from adolescence to young adulthood: I. Periods of risk
for initiation, continued use, and discontinuation.
American Journal of Public Health, 74, 660-666.
Kaplan, H. B., Martin, S. S., & Robbins, C. (1982).

Applications of a general theory of
deviant behavior: Self-derogation and adolescent drug use.

Journal of Health and

Katz, P., & Zigler, E. (1967). Self-image disparity: A
developmental approach.
Kazdin, A. E. (1979). Imagery elaboration and

self-efficacy in the covert modeling
treatment of unassertive behavior. Journal of Consulting

and Clinical Psychology,
47, 725-733.
King, A. C, Frey-Hewitt, B., Dreon, D. M., & Wood, P. D.

(1989). Diet vs. exercise in
weight maintenance. The effects of minimal intervention

strategies on long-term
outcomes in men. Archives of Internal Medicine,

149,2741-2746.
Koplan, J. P., Powell, K. E., Sikes, R. K., Shirley, R. W.,

Campbell, C. C (1982). An
epidemiologic study of the benefits and risks of running.

Journal of the American
Medical Association, 248, 3118-3121.
Levenson, R. W., & Gottman, 1. M. (1978). Toward the

assessment of social compe
tence. Journal of Consulting and Clinical Psychology,

46,453-462.
Marlatt, G. A., & Gordon, J. (1985). Relapse prevention:

Maintenance strategies in
addictive behavior change. New York: Guilford.
Miller, J. D., Cisin, I. H., Gardner-Keaton, H., Harrell,

A. V., Wirtz, P. W., Abelson,
H. I., & Fishburne, P. M. (1983). National survey on drug

abuse: Mainfindings 1982.
PHS, NIDA, DHHS Pub. No. (ADM) 83-1263. Washington, DC: U.

S. Government
Printing Office.
Newcomb, M. D. (1984). Sexual behavior, responsiveness,

and attitudes among
women: A test of two theories. Journal of Sex & Marital

Therapy, 10,272-286.
Newcomb, M. D. (1986). Nuclear attitudes and reactions:

Associations with depres
sion, drug use, and quality of life. Journal of

Personality and Social Psychology, 50,
906-920.
Newcomb, M. D., & Bentler, P. M. (1986). Loneliness and

social support: A confirma
tory hierarchical analysis. Personality and Social

Newcomb, M. D., & Bentler, P. M. (1987). Self-report

methods of assessing health
status and health service utilization: A hierarchical

confirmatory analysis. Multi
variate Behavioral Research, 22, 415-436.
Newcomb, M. D., & Bentler, P. M. (1988a). Consequences of

adolescent drug use:
Impact on the lives of young adults. Newbury Park, CA:

Sage.
Newcomb, M. D., & Bentler, P. M. (l988b). Impact of

adolescent drug use and social
support on problems of young adults: A longitudinal study.

Journal of Abnormal
Newcomb, M. D., & Harlow, L. L. (1986). Life events and

substance use among adole
scents: Mediating effects of perceived loss of control and

meaninglessness in life.
O'Leary, A. (1985). Self-efficacy and health. Behavior

Research and Therapy, 23,
437-451.
Ozer, E. M., & Bandura, A. (1990). Mechanisms governing

empowerment effects: A
self-efficacy analysis. Journal of Personality and Social

Pearlin, L., & Schooler, C (1978). The structure of

coping. Journal of Health and
Radloff, L. S. (1977). The CES-D scale: A self-report

depression scale for research in
the general population. Applied Psychological Measurement,

I, 385-401.
Roberts, R. E. (1980). Reliability of the CES-D scale in

different ethnic contexts.
Psychiatry Research, 2, 125-134.
Rothbaum, F., & Weisz, J. R. (1989). Child psychopathology

and the quest for control.
Newbury Park, CA: Sage.
Ryckman, R. M., Robbins, M. A., Thornton, B., & Cantrell,

P. (1982). Development
and validation of a physical self-efficacy scale. Journal

Shiffman, S., Read, L., Maltese, J., Rapkin, D., & Jarvik,
M. E. (1985). Preventing
relapse in ex-smokers: A self-management approach. In A.

Marlatt & 1. Gordon
(Eds.), Relapse prevention (Chapter 8, pp. 472-520). New

York: The Guilford Press.
Simpson, D. D., Joe, G. W., Lehman, W. E., & Sells, S. B.

(1986). Addiction careers:
Etiology, treatment, and 12-year follow-up outcomes.

Journal of Drug Issues, /6,
107-121.
Stall, R., & Biernacki, P. (1986). Spontaneous remission

from the problematic use of
substances: An inductive model derived from a comparative

analysis of the alcohol,
opiate, tobacco, and food/obesity literatures.

International Journal of the Addictions,
21,1-23.
Stein, 1. A., Newcomb, M. D., & Bentler, P. M. (1986).

Stability and change in per~0n
ality: A longitudinal study from early adolescence to young

adulthood. Journ.1! u}
Strecher, V. J., DeVellis, B. M., Becker, M. H., &

Rosenstock, 1. M. (1986). The 10k
of self-efficacy in achieving health behavior change.

Health Education Quartel/y, /3,73-91.
Taylor, S. E. (1986). Health psychology. New York: Random

House.
Weissman, M. M., Shokomskas, D., Pottenger, M., Prusoff,

B. A., & Locke. B. Z.
(1977). Assessing depressive symptoms in five psychiatric

populations: A validation
study. American Journal of Epidemiology, /06, 203-214.
Weitz, S. (1977). Sex roles. New York: Oxford University

Press.
Wood, R., & Bandura, A. (1989). Impact of conceptions of

ability on self-regulatory
mechanisms and complex decision making. Journal of

Personality and Social
Psychology, 56,407-415. Author Notes This research was

supported by grant DA01070 from the National Institute
on Drug Abuse. The computer assistance of Sandy Yu and

production assistance
of Julie Speckart are gratefully acknowledged.

EXPECTANCIES AS MEDIATORS BETWEEN RECIPIENT
CHARACTERISTICS AND SOCIAL SUPPORT INTENTIONS Ralf
Schwarzer, Christine Dunkel-Schetter, Bernard Weiner, and
Grace Woo It is assumed that the motivation to extend
social support is governed by specific emotions and
cognitions, among them outcome expectancies and
self-efficacy expectancies. Two experiments were conducted
to explore this assumption, Study I dealing with outcome
expectancy and Study /I dealing with self-efficacy
expectancy. In Study I, outcome expectancies toward eight
disease-related stigmas and the intention to extend social
support were examined with two experimental conditions.
The onset of the stigmas was varied as being either
controllable or uncontrollable. In addition, the target
person was described either as actively coping with the
stigma or as not actively coping. Examined were the
effects oj onset controllability and coping on pity,
outcome expectancy, and willingness to support the target
person. In a within-groups design, 84 subjects were
confronted with all eight stigmas under four different
conditions. Both experimental factors influenced the
reported reactions. The coping variable appeared to be
stronger than the controllability variable and. in
addition, outcome expectancy was a somewhat more important
mediator of helping than pity. However, the pattern of
data was context-specific, i.e., different sets of
predictors emerged for different stigmas. Study /I was a
similar experiment pursuing the notion that the motivation
to help is affected by the belief that one can be
effective as a helper (self-efficacy expectancy). It
examined whether self-efficacy expectancy for helping a
rape victim served as a mediator of the relationship
between recipient characteristics and support intentions.
The recipient characteristics assessed were victim coping
and controllability of the assault. Both pity and
self-efficacy expectancy emerged as good predictors of
support, whereas controllability and coping were of lesser
influence.
According to Bandura's cognitive-social theory, human
behaviors are partly
governed by expectancies, in particular by outcome

expectancies and self
efficacy expectancies (Bandura, 1977, 1986, 1991). Many

studies, some of them
presented in this volume, have applied this assumption to

specific behaviors in
various domains of human functioning such as achievement,

organizational man
agement, or health. There seems to be, however, no

application to studies on
social support. The willingness to help others depends

partly on one's emotions
at the time, but helping also depends on judgments about

the specific situation,
characteristics of the recipient, and one's self. Among

such cognitions are expect
ancies about the likelihood that the situation can be

changed and regarding one's
ability to provide the necessary social support. Expecting

a condition to improve
under certain circumstances represents an outcome

expectancy. Belief in oneself
as an effective support provider in a particular situation

represents a self-efficacy
expectancy. These cognitions are hypothesized to serve as

causal mediators of
the relationships between antecedent recipient

characteristics and consequent
intentions to extend social support. In addition, a number

of other factors outlined
below are considered important in the study of social

support. The present chapter reports two studies. The
first one deals with the mediat
ing role of outcome expectancy, the second one with the

mediating role of self
efficacy expectancy. In the following sections, we

describe in more detail the
constructs involved in this research, in particular

perceived controllability,
perceived coping, expectancies, and social support.

Perceived Controllability Attribution theory has recently
been extended to the study of social stigmas
and reactions to the stigmatized (Weiner, Perry, &

Magnusson, 1988). By social
stigma we mean a discrediting condition or mark that

defines a person as
"deviant, flawed, limited, spoiled, or generally

undesirable" (Jones et aI., 1984,
p. 6). Among others, physical deformities, behavioral

problems such as excessive
eating and drinking, and diseases can be regarded as

stigmas. Attribution theory
is relevant to the study of stigmas because individuals

typically search for the
cause(s) of a negative state or condition existing in

others. That is, observers con
fronted with a "markable" target initiate an attributional

search to determine the
origin of the stigma. Researchers have identified

controllability as one of the basic dimensions of
perceived causality (Weiner, 1985, 1986). Controllable

causes are those which an
actor can volitionally change, whereas uncontrollable

causes are not subject to
personal mastery or management. The onset of a drug

problem, for example, is
seen as controllable if a person has been experimenting

with drugs out of curi
osity, whereas it is perceived as comparatively

uncontrollable if a person has had
medical treatment with drugs and thereby developed a

dependency (Weiner et aI.,
1988). In a similar manner, the onset of a heart disease is

construed as control
lable if the person has led an unhealthy life-style,

including smoking and a poor
diet, whereas it is considered relatively uncontrollable

if hereditary factors have
played a major role in the illness. Affective and

Behavioral Reactions It has been documented that the
perceived controllability of a social stigma
determines disparate affective reactions toward the target

person and different
behavioral responses as well (e.g., Dejong, 1980; Weiner et

al., 1988). More
specifically, uncontrollable origins of stigmas tend to

elicit pity and offers of
help, whereas controllable origins tend to elicit anger and

no help (see
Reisenzein, 1986; Schmidt & Weiner, 1988; Weiner et al.,

1988). Hence, it has
been shown that experimentation with drugs and an unhealthy

life-style as causes
of stigmas yield much anger, little pity, and relative

neglect, whereas drug prob
lems due to medical treatment, and heart disease derived

from genetic factors,
give rise to little anger, much pity, and prosocial

responses. Perhaps more than in any other area within the
field of social motivation,
investigators of helping behavior have assumed that

emotions play an important
motivational role (see review in Carlson & Miller, 1987).

These emotions have
included discomfort (e.g., Cialdini, Darby, & Vincent,

1973), distress (e.g.,
Batson, O'Quinn, Fultz, Vanderplas, & Isen, 1983), empathy

(e.g., Batson, 1990;
Hoffman, 1975), gratitude (e.g., Goranson & Berkowitz,

1966), guilt (e.g.,
Hoffman, 1982), as well as pity and anger (Schwarzer &

Weiner, 1990, 1991).
While there is strong support for an

attribution-emotion-helping link, there are
also studies that have failed to demonstrate this effect.

Capitalizing on a real-life
event, Amato, Ho, and Partidge (1984) sent survey

questionnaires to residents
living near the setting of a major bushfire which killed

46 people and destroyed
over 2,000 homes. The questionnaire addressed perceptions

of causality and
responsibility, affective reactions, and helping behavior.

Most people reported
donating to the victims regardless of the amount of

responsibility attributed to
them. The obvious high degree of need in this context

seemed to have over
powered the attribution of control effects. Jung (1988)

presented subjects with vignettes depicting a close friend
experi
encing a variety of common problems, with manipulations of

the responsibility
for the problem. For each vignette, subjects rated the
target person's
deservedness of fate, perception of how helpful social

support would be for the
problem, and their likelihood of providing social support.

Perceived deservedness
of fate was greater for those viewed as having high

responsibility. Perceived
benefits of social support were also higher in this case.

However, neither factor
affected the likelihood of social support provision.

Skokan (1990) examined the affective responses and support
behaviors
extended towards a roommate who is dealing with either

cancer or the death of
her father. Subjects were presented with scenarios which

depicted the roommate
as either responsible or as not responsible for the onset

of the critical event. In
her initial within-subjects analysis, controllability was

associated with more
anger, less sympathy and less social support; however, when

reanalyzed as a
between-subjects design because of order effects, the

impact of controllability on
sympathy and support disappeared. Perceived Coping It

remains unclear whether stigma onset, which is a distant
event, is the sole
or main detenninant of affective and behavioral reactions

toward the stigmatized
or whether subsequent events, controllable or

uncontrollable, alter the causal
sequence. Drug experimentation and poor life-style, for

example, might be weak
predictors of the emotions and behaviors of observers when
compared with the
present efforts of the target person to cope with the

consequences of the stigma.
In the achievement domain, it is obvious that even after

failure due to lack of
effort, present expenditure of effort to compensate or

recover generates positive
affect and rewards for the failed student (Karasawa, 1991;

Weiner, 1985). When
generalized to the health domain, this finding suggests

that positive coping
attempts with a serious health condition could play an

important role in
detennining the affective and behavioral reactions of

others. Skokan (1990) distinguished in her scenario
experiment between adaptive
coping and maladaptive coping. In the adaptive condition,

the target person who
either had cancer or was bereaved, tried to stay

optimistic and to look for ways to
go on with her life and to grow from the experience. In the

maladaptive condi
tion, she dwelled on the negative aspects of the situation

and did not try to over
come the crisis instrumentally. Adaptive coping of the

target was related to less
anger in subjects but had mixed effects on their

willingness to offer social sup
port. In the bereavement condition, poor coping elicited

less support, but in the
cancer condition, unexpectedly, poor coping elicited even

more support. Silver, Wortman, and Crofton (1990) studied
subject reactions to a cancer
patient who was portrayed either as a "good coper," a "bad
coper" or a "balanced
coper." In the good coping condition, the target person

expressed an optimistic
view of her illness and appeared to be coping well. In the

balanced coping condi
tion, she conveyed distress about what was happening, but

also indicated that she
was trying her best. In the poor coping condition, she

displayed distress about
what was happening and appeared to have difficulty coping.

In nine out of ten
comparisons, the responses to confederates who were

portrayed as having posi
tive or balanced coping styles were significantly more

favorable than were
responses to poor copers. In sum, both the origin of a

problem and its solution are hypothesized to be
important when examining reactions of others toward the

stigmatized person
(Brickman et aI., 1982). That is, the responsibility for

causing a problem should
be separated from the responsibility for maintaining or not

alleviating it. This
important distinction has been ignored in prior research

on attributions (see also
Karasawa, 1991; Schwarzer & Weiner, 1991). The present

studies compare the
effects of perceived onset controllability with those of

perceived coping efforts
on pity, outcome expectancy, and social support towards the

stigmatized and
examines the mediating role of pity and expectancy.

Expectancies The focus of the present paper is on the
role of mediating factors that link
attributions and affect regarding a social stigma to

behavioral intentions or to
actual support behavior. Bandura (1977, 1986, t 991) has

convincingly demon
strated that expectancies are very important

social-cognitive mediators of action.
There are two major cognitions of this kind, outcome

expectancies and self
efficacy expectancies. In the first experiment, we deal

with outcome expectancies
that refer to the possibility of improvement of a

condition. The subjects were
asked how likely it is that a target person's condition

would improve under partic
ular circumstances. It is hypothesized that an

individual's active coping with an
ailment will trigger positive outcome expectancies in the

observer. Coping
behavior implicitly refers to the stability of a stigma.

If the victim is not actively
involved in alleviating the distress, maintaining

functioning and moving on with
daily life, one would have little reason to expect an

improvement; support may be
seen as wasted labor. If, however, a great deal of effort

is expended by the victim
in solving the problem, one can expect that changes are

more likely and that sup
plementary contributions would be a worthwhile investment.

This reasoning does
not apply to situations that require acceptance; that is,

we are likely to help
people who behave passively when passivity is required in
the situation. In the second experiment, the focus is on
self-efficacy expectancy in terms of
one's helping capabilities. Empathy, perspective taking,

comforting skills and so
on, not only facilitate social support in an objective

sense (Batson, 1990; Clary &
Orenstein, 1991); these abilities also have to be perceived

by the help provider in
order to establish a motivation to help. Help-specific

self-efficacy deals with cog
nitions about one's capability to support others and to

make a difference with this
support; it refers to one's perceived personal resources

to provide competent
assistance and to achieve relief for a sufferer. Social

Support Social support has been defined as an exchange of
resources "perceived by
the provider or the recipient to be intended to enhance the

well-being of the
recipient" (Shumaker & Brownell, t 984, p. t 3). This

definition requires that
either the provider or the recipient must perceive that the

provider has a positive
intent. Intentions have also been claimed as being the best

predictors of a variety
of behaviors; this is well-documented in research based

upon the Theory of
Reasoned Action (Fishbein & Ajzen, 1975) and the Theory of

Planned Behavior
(Ajzen, 1988). Evidence of the influence of help

intentions on actual helping
behavior has been found by Borgida, Simmons, Conner, and

Lombard (1990) and
Dalbert, Montada, and Schmitt (1988). Whether intentions to
help are accurately
perceived by the provider or by the recipient is a related

but different question
(Dunkel-Schetter & Bennett, 1990; Dunkel-Schetter,

Blasband, Feinstein, &
Bennett, 1991). Several factors determine the likelihood

that a supportive exchange actually
takes place. Stress factors, relationship factors,

recipient factors, and provider
factors have been discussed and somewhat studied

(Dunkel-Schetter & Skokan,
1990). We will deal here with the latter two exclusively.

Recipient factors are
critical determinants of support. Victims who are not only

distressed, but are also
not responsible for the event, and who invest a great deal
of effort to manage
their condition, are apt to elicit more help than those who
are responsible them
selves for their misfortune and who do not take action to

solve their issue
(Bennett-Herbert & Dunkel-Schetter, in press; Brickman et

aI., 1982). Creating
frustration and helplessness in the potential provider

leads to a lesser likelihood
of support (Dunkel-Schetter & Wortman, 1981, 1982). The

expression of too
much distress strains the social network, evokes negative

reactions, and turns
those away who would have been supportive if the distress

level had only been
moderate. Another reason why the network may not be

mobilized is if a victim is
not coping adaptively. Passive, depressive and ungrateful

victims or patients are
seen as socially unattractive and, therefore, receive less

support in the long run
(Barbee, 1990; Gurtman, 1986; Notarius & Herrick, 1988).

Paradoxically, those
subjects who have valuable personal resources such as

competence, high self
esteem, locus of control, and optimism and who make use of

their resources seem
to elicit a stronger tendency in others to extend support.

Provider factors have been intensively studied in social
psychology research
on helping (Batson, 1990; Berkowitz, 1987; Dovidio, 1984;

Eisenberg & Miller,
1987; Jung, 1988). It makes a difference how the cause of

the problem is attrib
uted. If it is seen as controllable then the victim is

blamed and negative emotional
reactions are aroused such as anger, leading to neglect of

the sufferer. If, on the
other hand, the cause is seen as uncontrollable and the

person does not seem to be
responsible for the problem, then positive emotions such as

pity emerge, which
make help more likely (Weiner, 1985). Thus, emotions are

mediators of attribu
tions and behavioral intentions. According to Batson

(1990), empathy predicts
altruistic motivation to help, whereas a provider's

distress tends to elicit egoistic
motivation, which does not induce help. These two theories

by Weiner and
Batson are closely related in terms of emotional mediators
of motivation. Pity can
be matched to empathy as a predictor of help, and anger

parallels distress in pre
dicting neglect. Betancourt (1990) has attempted to

integrate both views by
manipulating experimentally the controllability of onset

of a problem as well as
inducing different perspectives in the potential support

provider. He found that
both experimental factors influenced perceived

controllability and empathic
emotions that, in tum, influenced helping. In the present

chapter, the focus is on experimentally manipulated
recipient
factors, but it is kept in mind that these do not operate

in an isolated manner.
Rather, they interact with on-going responses by the

provider during a specific
social encounter. It is only of secondary importance

whether the victim is
actually responsible for the problem and whether active

coping is executed.
Moreover, the degree to which the provider makes these

attributions, is
considered to be critical. The perception in the beholder

may be more relevant
than the actual cause of the onset of the stigma or the

actual coping behavior. The present experiments were
designed to examine the effects of perceived
controllability and perceived coping on pity as an

affective reaction and on
expectancies that, in turn, were hypothesized to exert an

influence on support
intent. The studies differ in terms of the scenarios used
and in terms of the
expectancy variables. While Study I deals with outcome

expectancy, Study II
deals with self-efficacy expectancy. STUDY I Method

Sample. The subjects were 84 male and female students at
the University of
California, Los Angeles, who received credit in an

introductory psychology
course for their participation. They were randomly assigned

to one of four groups
(see below) and given questionnaires in small group

sessions with anonymity
assured. I Design. Eight health-related stigmas were

selected, each of which was
manipulated with respect to onset controllability and

coping effort. Each subject
received four of the eight stigmas paired uniquely with one

of the four control
lability conditions (2 Levels of Onset Responsibility x 2

Levels of Coping). Sub
jects were divided into four groups that received different

combinations of
stigmas and conditions (see Table I).
Table I
Experimental Design Onset Responsible Onsel Irresponsible

No Coping Coping No Coping Coping
Group 1 Aids Cancer Drug abuse Heart disease
Group 2 Cancer Drug abuse Heart disease Aids
Group 3 Drug abuse Heart disease Aids Cancer
Group 4 Heart disease Aids Cancer Drug abuse

Group 1 Anorexia Child abuse Depression Obesity
Group 2 Child abuse Depression Obesity Anorexia
Group 3 Depression Obesity Anorexia Child abuse
Group 4 Obesity Anorexia Child abuse Depression
1 Study I was conducted by Ralf Schwarzer and Bernard

Weiner. As shown in Table 1, one part of the design
included four stigmas (AIDS,
cancer, drug abuse, and heart disease) paired with the four
conditions, while a
second part replicated the flrst but used another four

stigmas (anorexia, depres
sion, obesity, and child abuse). Thus, there were two

within-group factors (onset
controllability and coping) and one between-group factor

(stigma set). This
design allowed for an overall analysis as well as for

stigma-specific subanalyses. Four vignettes were created
for each stigma consisting of: (a) onset
responsibility and low coping; (b) onset responsibility and

high coping; (c) no
onset responsibility and low coping; and (d) no onset

responsibility and high
coping. As an example, the obesity vignettes are given:
1. Maladaptive coping, controllable. ¥our roommate has

become excessively overweight, and is experiencing severe
problems in socialand work-related activities. Excessive
eating and lack of exercise have been the primary
contributors to the obesity. This roommate does not take
any steps to lose weight, either by dieting, exercising
or by following a medical regimen.
2. Adaptive coping, controllable. Your roommate has become

excessively overweight, and is experiencing severe
problems in socialand work-related activities. Excessive
eating and lack of exercise have been the primary
contributors to the obesity. Recently this roommate has
commenced a new diet prescribed by a physician, and is
regularly exercising.
3. Maladaptive coping, uncontrollable. ¥our roommate has

problems in socialand work-related activities. Glandular
dysfunction has been identified as the reason for the
obesity. This roommate does not take any steps to lose
weight, either by dieting, exercising or by following a
medical regimen.
4. Adaptive coping, uncontrollable. Your roommate has

problems in socialand work-related activities. Glandular
dysfunction has been identified as the reason for the
obesity. Recently this roommate has commenced a new diet
prescribed by a physician. and is regularly exercising.
Measures. The dependent variables were the following
9-point rating scales,
anchored with extremes such as not at all and very much

so. Pity was assessed by
the single item "How much pity would you feel?"

Typically, outcome expectancies are worded in an "if-then
manner." In the
present experiment, however, the if-component was given by

the four experi
mental conditions such as: "If the stigma is

uncontrollable and if the victim is
actively coping with it, then ... " Because of these

implicit assumptions, the mea
surement of the outcome expectancy was restricted to the

then-component and
simply worded: "How likely is it that the condition will

improve?" Social support intention was measured by seven
items representing different
kinds of social support. However, this was a homogeneous

scale (Cronbach's
alpha for the seven social support items was .91), and,
therefore, the aggregated
score was used as an indicator of support intentions. The

items were: 1. How much would you like to extend support
to your roommate? 2. How much time would you be willing to
spend talking and listening? 3. How much money would you
be willing to donate in order to provide the best
possible treatment? 4. How much would you like to go on a
holiday trip with your roommate? 5. How much would you be
willing to give advice and information? 6. How much would
you be willing to console and reassure your roommate when
being upset? 7. How willing would you be to assist with a
small problem? Other dependent variables were analyzed
previously within the framework of
analysis of variance, and some of the results are published

elsewhere. However,
we have only reported about the stigmas of heart disease

(Schwarzer & Weiner,
1990), AIDS and cancer (Schwarzer & Weiner, 1991). Results

To examine the role of pity and outcome expectancy as
mediators of the rela
tionship between victim characteristics and provider

support intentions, a
structural equation model was specified with

controllability and coping as exo
genous variables and pity, expectancy, and support as

endogenous variables. This
is a straightforward single indicator model with manifest

variables. The two
orthogonal experimental factors were believed to influence

emotions and cogni
tions, whereas emotions and cognitions were specified to

influence the behavioral
intention directly. Controllability and coping, therefore,

could exert indirect
effects on support intent through pity and expectancy but

were constrained not to
exert direct effects, because this would not be in line

with theory or past research.
The two alternative mediating factors were pity and

expectancy, and for both of
them the size of their mediating effect was computed in
addition to their direct
impact on support intent (see Figures 2 to 9). This

procedure was repeated eight
times, for each stigma individually. Eight path analyses

were carried out with the
LISREL VII program (Joreskog & Sorbom, 1988). First, the

degree to which the experimental data fitted the structural
equation
model was examined. Several indices of fit have been

suggested in the literature
(cf. Bentler, 1980). We have used five of them in this

study, (a) the chi-square
test which, if significant, indicates that the data

deviate from the model, (b) the
chi-square Idf ratio which takes the degrees of freedom

into account (df = 3) and
which should be as low as possible; ratios above 3.0 are

usually seen as unsatis
factory, (c) Joreskog's Goodness of Fit Index (GFl) which

should be close to
unity, (d) his Adjusted Goodness of Fit Index (AGFl) that

makes an adjustment
to the degrees of freedom and also should be as high as

possible, and (e) the Root
Mean Square Residual (RMSR) which is an index derived from

the deviations of
the original correlation matrix from the reproduced

correlation matrix on the
basis of the estimated parameters; this index should not

exceed .05. Table 2 summarizes the results of all eight
path analyses. In six of eight
cases, an excellent fit emerged, whereas the stigmas

"Cancer" and "Child Abuse"
turned out to be associated with a less appropriate fit.
Overall, these satisfactory
results indicate that the model specification is in line

with the experimental data,
but also that the specific stigma context makes a

difference.
Table 2
Goodness of Fit for the Eight Path Models
Stigma chi2 p chi 2 /df OF! AOFI RMSR
AIDS 4.36 .23 1.45 .98 .90 .05
Cancer 9.49 .02 3.16 .96 .79 .07
Drug abuse 1.46 .69 0.49 .99 .97 .03
Heart disease 2.44 .49 0.81 .99 .94 .04
Anorexia 0.92 .82 0.31 .99 .98 .02
Child abuse 15.18 .002 5.06 .94 .69 .08
Depression 1.50 .68 0.50 .99 .97 .03
Obesity 4.65 .20 1.55 .98 .90 .05
Note. OFI == goodness of fit, AOFI == adjusted OFI, RMSR ==

root mean square residual.
Table 3
Percent of Explained Variance Endogenous Factor Stigma

Pity Expectancy Support AIDS 22 I 31 Cancer 6 29 21
Drug abuse 9 51 15 Heart disease 1 31 11 Anorexia 4
51 6 Child abuse 9 36 29 Depression 16 14 Obesity
38 11 This is corroborated by the explained variance for
the three endogenous vari
ables pity, expectancy and support (Table 3). The model

succeeded in explaining
a great deal of the variance of expectancy and support but

much less so of pity.
This shows that the emotion of pity is not sufficiently
predicted by controllability
and coping. Other factors, not under scrutiny here, must

be responsible for the
variation in pity. The stigma-specific path coefficients

are displayed in Figures 1-8; Table 4
contains the decomposition of total effects into direct

and indirect effects. Results
for each stigma will be described briefly. Coefficients

above .21 are significant.
Table 4
Decomposition of Effects on Social Support Intention
Stigma Predictor Direct Indirect Total Effect
AIDS Control 0 21 21 Coping 0 14 14 Pity 55 0 55

Expectancy 7 0 7
Cancer Control 0 6 6 Coping 0 13 13 Pity 43 0 43

Expectancy 7 0 7
Drug abuse Control 0 6 6 Coping 0 29 29 Pity 14 0 14

Expectancy 40 0 40
Heart disease Control 0 3 3 Coping 0 16 16 Pity 19

0 19 Expectancy 26 0 26
Anorexia Control 0 0 0 Coping 0 17 17 Pity 14 0

14 Expectancy 24 0 24
Child abuse Control 0 12 12 Coping 0 25 25 Pity 37 0

Depression Control 0 4 4 Coping 0 14 14 Pity 14 0

Obesity Control 0 4 4 Coping 0 18 18 Pity 14 0 14

Expectancy 30 0 30 In the case of AlDS, a substantial
causal path leads from controllability to
pity (p = -.39) and another from pity to support (p = .55).

Coping has a somewhat
lower impact on pity (p = .25). Expectancy does not playa

role: it is predicted
neither by controllability nor by coping, and it does not

predict support. Since
AIDS is a terminal disease, it is not surprising not to

find a large variation in out
come expectancy. Pity appears to be the appropriate

emotional reaction which
facilitates the likelihood to extend support (see Figure t

and Table 4) .
Figure I Pity and expectancy as mediators between

controllability and coping and social support in the AIDS
scenario. For cancer, pity was again the best predictor of
support (p = .43), whereas
expectancy failed to contribute anything (p = .07). But the

antecedents were dif
ferent; controllability had no significant impact on pity

or expectancy, whereas
coping had a strong path to expectancy (p = .52) and a

moderate one to pity (p =
.22). Although cancer can be a terminal disease in many

cases, there are better
survival chances for those who comply with treatment. This

explains the associa
tion between coping and expectancy, but, surprisingly,

there was little effect on
support intentions that were based more on pity (see Figure

2 and Table 4) .
Figure 2 Pity and expectancy as mediators between

controllability and coping and social support in the
cancer scenario. In case of drug abuse, variations in
controllability elicited no effects reac
tions but coping did. A strong path from coping to

expectancy emerged (p = .70), Control -.39 Pity .55 .25
Support .06 .08 Coping .07 -.12 Control Pity .43 .2 2 .13
Coping .52 Expect .07 Support Expect
accompanied by another strong path from expectancy to
support (p = AD) making
this the major pathway to help intentions. A minor pathway

was added from cop
ing through pity (p = .25, P = .13). Drug abuse is a

rather unstable condition and
appears to be modifiable. Whether one is ready to support a

drug user mainly
depends on the likelihood of perceived change based on his

or her coping efforts.
no matter how the problem was originally caused (see Figure

3 and Table 4) .

controllability and coping and social support in the drug
abuse scenario. In case of heart disease, there was no
effect of controllability, and pity also
had no significant relationships (see Figure 4 and Table 4.

The only pathway to
support led from coping via expectancy (p = .56, p = .26).

Heart disease is inter
preted as a modifiable condition that varies with one's

health behavior such as
nutrition, exercise, and relaxation. The origin of this

condition seems to be unim
portant for a decision to help the patient.

heart disease scenario. Anorexia nervosa can also be
regarded as an unstable condition where active
coping makes a difference. Controllability had no influence

but coping deter
mined expectancy (p = .70) and pity (p = .20) (see Figure

5 and Table 4). -.15 Control Pity .25 .09 .7 0 Coping
Expect .40 .13 Support Control -.04 Pity .19 .06 Support
.26 Expect .56 Coping .07
Anorexia is considered a highly modifiable condition. If a
patient copes well it
will vanish, no matter how controllable the origin was .

anorexia scenario. A different picture emerged for child
abuse. Both direct effects on support
were almost equal, with pity (e = .37) and expectancy (e =

.31) accounting for a
similar amount of variation in support. The key antecedent

factor, however, was
coping which was closely related to expectancy (p = .60).

Compared to drug
abuse, child abuse is not a health-compromising behavior

but more a socially
deviant act that elicits emotions such as either outrage or

pity towards the actor,
the latter emotion only if there was not much control over
the behavior (see
Figure 6 and Table 4).

child abuse scenario. In case of depression, the
predictors controllability and pity turned out to be
irrelevant, whereas expectancy had an influence on support

(p = .34), based on
the coping efforts of the target person (p = .38). This

clearly documents that an
active contribution on behalf of the mental health patient

is required in order to
make the condition look changeable, so that support would

not be in vain. Only
expectancy had an effect on support (see Figure 7 and Table

4). Control -.01 Pity .20 .14 Support .24 .12 Expect .70
Coping Control -.25 Pity .37 .18 .31 Support .10 Copin g
.60 Expect Finally, in case of obesity, an almost identical
result emerged. Again, control
lability and pity were negligible factors but expectancy (p

= .30), based on
coping (p = .61), made the difference. Obesity is an

unstable condition, and those
who do not counteract their problem cannot count on help

from others. Only
active coping efforts elicit expectancy which in tum

trigger readiness for social
support (see Figure 8 and Table 4).

depression scenario.

obesity scenario. In sum, in five of the eight stigmas,
outcome expectancy was the main pre
dictor of support intention. These five were drug abuse,

heart disease, anorexia,
depression, and obesity. The two terminal diseases, AIDS

and cancer, differed
from the majority by their conspicuous pathway from pity to

support intent. In
these two cases, one's intention to help was almost

exclusively based on pity. For
child abuse, a balanced influence of pity and expectancy

emerged. Coping was a
stronger antecedent than controllability in seven out of

eight cases. The exception
was AIDS. The overall picture corroborates the assumption

that outcome
expectancy is a critical mediator between target coping and

social support inten
tion. From these results, whether one extends help or not
is primarily dependent
on the expectancies aroused by the victim characteristics,

and particularly the
person's way of coping. -.01 Control Pity .14 Support .08

.12 Coping .38 Expect .34 -.09 Pity Control .14 .0 4 .09
Coping .61 Expect .30 Support Discussion Each of the eight
stigmas was examined in separate path analyses with
respect to the two experimental factors, controllability

and coping as antecedents,
and pity and expectancy as mediators. The model fit the

data and expectancies
were a major direct source of support variation. Pity was

a direct predictor of
social support only in three specific contexts. It is

noteworthy that there was a
high degree of variation between the eight stigmas,

indicating that the specific
circumstances decide whether the willingness to help is

primarily based on either
pity or expectancy. In terminal diseases such as AIDS or

cancer, pity appeared to
be more influential than expectancy, whereas for unstable

health conditions such
as drug abuse, anorexia or obesity the

coping-expectancy-support link was
obvious. It might be, therefore, that the perceived

stability of a condition is a crit
ical underlying dimension that affects judgments of help.

Controllability was less
influential compared to coping which, in turn, partly

determined expectancy. The
most conspicuous pathway led from coping via expectancy to

support intent. STUDY II In the first experiment, the
expected improvement of the target's condition
was one of the mediators under investigation. In the
second experiment, the
attention was shifted to a support provider characteristic

to address the question
of whether the perception of one's ability to help would

make it more likely that
a support intention occurs. In other words, self-efficacy

expectancy, one's per
ceived personal capability of extending effective support,

was the focus. It was
hypothesized that self-efficacy expectancy played the same

role as a mediator
that outcome expectancy did in the first experiment. 2

Method Design. The path-analytic model was the same as in
the first study but there
were some differences in the experimental manipulations and

in the measures in
volved. Only one problem situation was selected, a sexual

assault scenario, that
was varied with respect to controllability and coping. A

rape victim in the
uncontrollable condition was described as a student who

studied one night at the
library and was raped on the way to her car by a stranger.

In the condition
designed to seem slightly more controllable she was

described as someone who
attended a party where she drank too much and flirted with
the males; when she
was taken home by one of them, she invited him up to her

apartment and was
raped. The adaptively coping victim was characterized as

one who was trying
hard to go on with her life after the assault, having
joined a support group and
seeing a counselor each week. The maladaptively coping

victim did not try to
2 Study II was conducted by Grace Woo, Christine

Dunkel-Schetter, and Ralf
Schwarzer.
overcome her problem situation. She had withdrawn from

friends and did not eat;
she also refused to attend a support group meeting and to

see a counselor. The experiment was arranged as a 2 x 2
between-subjects design; 70 under
graduate students responded to the vignette randomJy

assigned to one of four
conditions. There were 55 males and only 15 females, but

their distribution over
the four cells was about equal, with cell sizes of 19, 17,
18 and 16. Measures. Pity, self-efficacy expectancy and
support intentions were the
dependent variables used in this report. All were rated on

a 5-point scale. Pity
was assessed by four adjectives as part of a checklist,

namely empathy, sym
pathy, pity, and compassion. Emotional support intent was

measured by four
items such as "Would you be willing to try to console and

reassure your friend
when she is upset?" and "Would you spend time listening to

her emotional reac
tions to the assault?" Tangible support intent was

measured by six items such as
"Would you be willing to offer her help with her school

work if she needed it?"
and "Would you lend her money to see a therapist?"

Self-efficacy expectancy
was measured by a newly developed 10-item scale that was

employed for the first
time. Its psychometric properties were satisfactory with an

average item-total
correlation of .55 and an internal consistency of

Cronbach's alpha = .85. The
items were worded in the following way:
I. I possess the necessary social skills to alleviate the

distress of a sexual assault victim.
2. It is easy for me to comfort someone in distress.
3. I am capable of providing the appropriate resources for

a rape victim.
4. It is difficult for me to communicate empathic

understanding. (-)
5. I could make someone feel better no matter how depressed

she is.
6. When it comes to comforting someone, I feel awkward. (-)
7. I am not sensitive enough to meet the support needs of a

sufferer. (-)
8. I do not trust my skills to communicate in a beneficial

way with a sexual assault victim. (-)
9. I am not the kind of person who can meet the emotional

needs of others who are in a crisis. (-)
10. I have sufficient communication skills to cheer up

someone who is experiencing stress. Results A structural
model was specified with the two experimental factors as
ante
cedents, and with pity, self-efficacy expectancy, and

support as the dependent
variables. In contrast to the previous study, this is a

multiple indicator model. The
three endogenous variables were specified with two

indicators each. The four
pity items were divided into two sets (each pity indicator
had two items); support
was specified by the emotional support scale as well as the

tangible support scale,
and the two self-efficacy indicators were two 5-item

subsets of the instrument
described above. The results of the LlSREL analysis are

depicted in Figure 9.

controllability and coping and social support in the rape
scenario. The fit of the model was chi-square = 17.4 (15
df, p = .295) with a chi
square/df ratio of 1.16. Goodness of fit was GFI = .94 and

adjusted goodness of
fit AGFI = .87. The root mean square residual was RMSR =
.09. Although the
latter two indices fall short of the usual requirement,

the overall fit can be regard
ed as satisfactory, based on the other indices. The

explained variance for social
support was 34%, which is quite good. whereas those for

pity and for self
efficacy expectancy were low (7% each). Decomposing the

effects on support led
to substantial direct effects for pity (e = .44) and for

self-efficacy expectancy (e =
.36), and to smaller indirect effects for controllability

(e = .03). and for coping
(e = .17). Pity and expectancy were very good predictors

of support intent, but
the underlying experimental factors (control, coping) were

of lesser influence. Discussion The second experiment has
replicated the general causal model leading from
victim characteristics to support provider emotional
reactions or cognitions,
resulting in an intention formation. Pity emerged here as

the strongest predictor
of support, but self-efficacy expectancy also contributed

substantially. Controlla
bility turned out to be negligible, whereas coping exerted

a weak. but statistically
significant. influence on pity and self-efficacy. However.

it is difficult to construe a sexual assault as
controllable, and the
two conditions differed in ratings of controllability only

by one point, although
significantly. In addition. the rape scenario is quite

different from the eight
stigma scenarios described in Study I. There is no disease

or bodily condition Control -.12 .86 .70 Pity .23 44 Copin
g .06 Self-eff. Expectancy .25 .62 95 .36 Support .90 80
involved but a single violent act caused by an external

agent. An assault is likely
to be viewed generally as less controllable than other

social stigmas such as
obesity or drug abuse. The degree of controllability only

varied in the study from
uncontrollable to somewhat controllable; there was not

really a "controllable"
experimental condition. Adverse chance events seem

especially likely to trigger
pity, whereupon the victim is not blamed. GENERAL

DISCUSSION The present findings from Studies I and II are
based on hypothetical
scenarios with students. Therefore, the results can be

generalized neither to actual
helping situations nor to other populations. This procedure

also has some inher
ent limitations in that respondents may be unable to judge
accurately their affec
tive reactions and whether they would or would not offer

help to particular indi
viduals. In addition, some key variables that affect

emotion and social support
certainly are excluded from the manipulated factors.

However, as noted by
Cooper (1976), "when looked at from the point of view of

generating hypotheses,
finding new leads, and initiating models of behavior,

lrole playing I may be the
rbestl method" (p. 605). In addition, in the

investigations presented here and by
Weiner et al. (1988), the stimulus configurations examined

could not be found
without overwhelming difficulty in field research, with the

consequences that
variables would be confounded. Finally, prior research has

suggested that role
enactment strategies in the study of help-giving have

yielded data comparable
with observations of actual behavior (see review in

Weiner, 1986). For these
reasons, and particularly in light of the relatively

recent growth of the study of
social support, we used a hypothetical scenario method.

Research must extend
theoretical and experimental analyses within the current

framework before
applying these research questions to real-life situations.

The present studies have underscored the notion of
emotional and cognitive
mediators in the process of forming behavioral intentions.
When dealing with
victims of life events including medical patients, the

likelihood of mobilizing
help is dependent on a number of recipient and provider

characteristics (Dunkel
Schetter & Skokan, 1990). The controllability of the cause

of the problem ap
pears to playa role in the determination of help.

Moreover, the changeability or
instability of the problem as reflected in coping efforts

seems to elicit positive
expectancies in the observer and motivation to help. Such

efforts may create both
a sense that the situation can be improved and a belief

that one can effectively as
sist the victim. Thus, outcome expectancy as well as

self-efficacy expectancy are
useful cognitive mediators. They are part of a mechanism

that governs the trans
lation of thought into action. Both studies have dealt

with one of these cognitions
exclusively, and it would be worthwhile to integrate both

concepts into one
empirical framework in a subsequent study. One conclusion

of the first experiment concerns the specificity of the
result
pattern to individual situations. To what degree pity or

expectancy mediate
reCIpIent characteristics and support intent depended on

the particular
circumstances, i.e., the stigma chosen and, probably, the

unique wording of the
vignettes. In the second experiment, there was only one

context provided, namely
the rape scenario. Therefore, it remains unclear, as to

whether these
circumstances have affected the results. It could be, for

example, that for a
divorce or an accident, completely different path

coefficients would emerge. The
evidence for self-efficacy expectancy as a mediator is

limited to the context
chosen, and further research should make use of a number

of different problem
domains. There are underlying similarities, however,

between the selected problems
that may suggest a common pattern of reactions to victims.

For AIDS, cancer,
and rape, the emotion of pity appears to be a stronger

mediator than expectancy.
These problems are loss/harm situations, whereas contexts

such as anorexia,
obesity, drug abuse, child abuse, depression and heart

disease are more like
threats (see Dunkel-Schetter et aI., 1991; Hobfoll, 1988;

Lazarus, 1990; Lazarus
& Folkman, 1984). Different stress appraisals may determine

the amount of pity
and specific expectancies in potential support providers.

If a victim is severely
harmed or if the physical integrity of a victim is lost,

then pity prevails; if, how
ever, an on-going risky or threatening behavior is the

topic, it is seen as more un
stable and modifiable and, therefore, gives rise to a

greater role for expectancies. Expectancies can be
pessimistic or optimistic. Pessimism undermines the
motivation to help because the investment of further
support efforts appears to be
wasted; optimism, however, assumes that the victim will be

responsive to future
support attempts and thereby render them worthwhile.

Optimism, as a psycholog
ical construct, has been defined as "generalized outcome

expectancies" (Scheier
& Carver, 1985, 1987). This construct has recently become

one of the key issues
in research on stress, coping, and mental health as well as

physical health
(Scheier et aI., 1989; Seligman, 1991). The present studies

have underscored the
role of situation-specific outcome expectancies and

self-efficacy expectancies
after Bandura (1977, 1991). Further research should address

the notion of specifi
city versus generality of expectancy, with dispositional

optimism being one
example of a more general construct. Jerusalem and

Schwarzer (this volume)
have developed a global self-efficacy scale that has

demonstrated high predictive
and construct validity in several field studies. Although

specific measures are
preferred in clinical intervention studies of behavioral

change, there might be an
advantage to global measures in other research domaiils.

Although the present studies have provided preliminary
evidence for the role
of expectancies as mediators in the helping process, it

remains undetermined how
outcome expectancy and self-efficacy expectancy are
interrelated. Each experi
ment has dealt with only one of these cognitions but

failed to account for their
joint influence. It would be premature to conclude from

the above findings that
outcome expectancy exerts a stronger influence on support

intent than self
efficacy expectancy. There might be a causal order among

the two. For example,
it might be that a support provider does not scrutinize her

helping capability
unless being faced with a target's condition that is

improving or one that is, at
least, modifiable. A third variable could be critical here,

namely one's personal
experience with (a) crisis situations that require support,

and with (b) the effec
tiveness of one's previous helping attempts

(Dunkel-Schetter & Skokan, 1990).
Self-efficacy expectancy is shaped by context-specific

mastery experiences,
among others, and therefore it would be necessary to

investigate expectancies
Ajzen, I. (1988). Attitudes, personality. and behavior.

Milton Keynes: Open University
Press.
Amato, P. R. Ho, R, & Partidge, S. (1984). Responsibility

attribution and helping
behavior in the Ash Wednesday bushfires. Australian

Journal of Psychology, 36, 191-203.
Bandura, A. (1977). Self-efficacy: Toward a unifying

theory of behavioral change.

action. Englewood Cliffs, NJ:
Prentice-Hall.
Bandura, A. (1991). Self-efficacy conception of anxiety.

[n R. Schwarzer &
R Wicklund (Eds.), Anxiety and selflocused attention (pp.

89-110). London:
Harwood.
Barbee, A. P. (1990). Interactive coping: The cheering-up

process in close relationships.
In S. Duck with R Silver (Ed.), Social support and

personal relationships (pp. 46
65). London: Sage.
Batson, C. D. (1990). How social an animal? American

Psychologist, 45,336-346.
Batson, C. D., O'Quinn, K., Fultz, J., Vanderplas, M., &

Isen, A. (1983). Influence of
self-report distress and empathy on egoistic versus

altruistic motivation to help.
Bennett-Herbert, T., & Dunkel-Schetter, C. (in press).

Negative social reactions to
victims: An overview of responses and their determinants.

In L. Montada, S. H.
Filipp, & M. J. Lerner (Eds.), Life crises and experiences

of loss in adulthood.
Bentler, P. M. (1980). Multivariate analysis with latent

variables: Causal modeling.

Berkowitz, L. (1987). Mood, self-awareness, and
willingness to help. Journal of
Betancourt, H. (1990). An attribution-empathy model of

helping behavior: Behavioral intentions and judgments of
help-giving. Personality and Social Psychology Bulletin,
16,573-591.
Borgida, E., Simmons, R. G., Conner, C., & Lombard, K.

(1990). The Minnesota Living
Donor Studies: Implications for organ procurement. In J.

Shanteau &. R. Harris (Eds.), Psychological research on
organ donation (pp. 108-121). Washington, DC:
American Psychological Association.
Brickman, P., Rabinowitz, P. c., Karuza, J. Jr., Coates,

D., Cohn, E., & Kidder, L. (1982). Models of helping and
coping. American Psychologist, 37, 368-384.
Carlson, M., & Miller, N. (1987). Explanation of the

relation between negative mood and helping. Psychological
Bulletin, 102,72-90.
Cialdini, R., Darby, B., & Vincent, J. (1973).

Transgression and altruism: A case for hedonism. Journal
of Experimental Social Psychology, 9, 502-516.
Clary, E. G., & Orenstein, L. (1991). The amount and

effectiveness of help: The relationship of motives and
abilities to helping behavior. Personality and Social
Psychology Bulletin, 17,58-64 .
. Cooper, J. (1976). Deception and role playing. American

Psychologist, 31, 605-610.
Dalbert, c., Montada, L., & Schmitt, M. (1988). Intention

and ability as predictors of change in adult daughters'
prosocial behavior towards their mothers. Verantwortung.
Gerechtigkeit. Moral, 46, 1-17.
Dejong, W. (1980). The stigma of obesity: The consequences

of naive assumptions concerning the causes of physical
deviance. Journal of Health and Social Behavior, 21,75-87.
Dovidio, J. F. (1984). Helping behavior and altruism: An

empirical and conceptual overview. In L. Berkowitz (Ed.),
Advances in experimental social psychology (Vol. 17, pp.
362-428). New York: Academic Press.
Dunkel-Schetter, c., & Bennett, T. L. (1990).

Differentiating the cognitive and behavioral aspects of
social support. In I. G. Sarason, B. R. Sarason, & G. R.
Pierce (Eds.), Social support: An interactional view (pp.
267-296). New York: Wiley.
Dunkel-Schetter. C., Blasband, D. E., Feinstein, L. G., &

Bennett, T. L. (1991). Elements of supportive social
interactions: When are support attempts effective? In S.
Spacapan & S. Oskamp (Eds.), Helping and being helped in
the real world (pp. 83114). Newbury Park, CA: Sage.
Dunkel-Schetter, C., Folkman, S., & Lazarus. R. S. (1987).

Correlates of social support receipt. Journal of
Personality and Social Psychology, 53. 71-80.
Dunkel-Schetter, c., & Skokan, L. A. (1990). Determinants

of social support provision in personal relationships.
Journal oj Social and Personal Relationships. 7, 437-450.
Dunkel-Schetter, c., & Wortman, C. B. (1981). Dilemmas of

social support: Parallels between victimization and aging.
In I. B. Kiesler, 1. N. Morgan, & V. K. Oppenheimer (Eds.),
Aging: Social change (pp. 349381). New York: Academic
Press.
Dunkel-Schetter, C., & Wortman, C. B. (1982). The

interpersonal dynamics of cancer: Problems in social
relationships and their impact on the patient. In H. S.
Friedman & M. R. DiMatteo (Eds.). Interpersonal issues in
health care (pp. 69-100). New York: Academic Press.
Eisenberg, N., & Miller. P. A. (1987). The relation of

empathy to prosocial and related behaviors. Psychological
Bulletin, 101,91-119.
Fishbein, M., & Ajzen, 1. (1975). Belief, attitude.

intention. and behavior: An introduction to theory and
research. Reading, MA: Addison-Wesley.
Goranson, R. E., & Berkowitz, L. (1966). Reciprocity and

responsibility reactions to prior help. Journal of
Gurtman, M. B. (1986). Depression and the response of

others: Re-evaluating the evaluation. Journal of Abnormal
Hobfoll, S. E. (1988). The ecology of stress. Washington,
DC: Hemisphere.
Hoffman, M. L. (1975). Developmental synthesis of affect

and cognition and its implications for altruistic
motivation. Developmental Psychology, II, 607-622.
Hoffman, M. L. {I 982). Development of pro social

motivation: Empathy and guilt. In N. Eisenberg (Ed.),
Development oj prosocial behavior (pp. 281-313). New York:
Academic Press.
Joreskog, K., & Sorbom, D. (l988). LlSREL VII: A guide to

the program and applications. Mooresville, IN: Scientific
Software.
lones, E. E., Farina, A., Hastorf, A. H., Markus, H.,

Miller, D., & Scott, R. A. (1984). Social stigma. The
psychology of marked relationships. New York: Freeman.
lung, l. (1988). Social support providers: Why do they

help? Basic and Applied
Karasawa, K. (1991). The effects of onset and offset

responsibility on affects and helping judgments. Journal
of Applied Social Psychology, 21,482-499.
Lazarus, R. S. (1990). Theory-based stress measurement.

Psychological Inquiry, 1,3-13.

Springer.
Notarius, C. I., & Herrick, L. R. (1988). Listener

response strategies to a distressed
other. Journal of Social and Personal Relationships,

5,97108.
Reisenzein, R. (1986). A structural equation analysis of

Weiner's attribution-affect
model of helping behavior. Journal of Personality and

Social Psychology, 50, 11231133.
Scheier, M. F., & Carver, C. S. (1985). Optimism, coping,
and health: Assessment and
implications of generalized outcome expectancies. Health

Scheier, M. F., & Carver, C. S. (1987). Dispositional

optimism and physical well-being:
The influence of generalized outcome expectancies on

health. Journal of Personality,
55,169-210.
Scheier, M. F., Matthews, K. A., Owens, 1., Magovern, G. 1.

Sr., Lefebre, R. c., Abbott,
R. A., & Carver, C. S. (1989). Dispositional optimism and

recovery from coronary
artery bypass surgery: The beneficial effects on physical

and psychological well
being. Journal of Personality and Social Psychology.
Schmidt, G., & Weiner, B. (1988). An

attribution-affect-action theory of motivated
behavior: Replications examining judgments of help-giving.

Schwarzer, R., & Leppin, A. (1991). Social support and

health: A theoretical and
empirical overview. Journal of Social and Personal

Relationships, 8,99-127.
Schwarzer, R., & Weiner, B. (\ 990). Die Wirkung von

Kontrollierbarkeit und Bewaltigungsverhalten auf Emotionen
und soziale Unterstiitzung rThe effect of
controllability and coping on emotions and social supportl.

Zeitsrhrift fur Sozialpsy·
chologie, 2/,118-125.
Schwarzer, R., & Weiner, B. (1991). Stigma controllability

and coping as predictors of
emotions and social support. Journal of Social and Personal
Relationships, 8, 133140.
Seligman, M. E. P. (1991). Learned optimism. New York:

Knopf.
Shumaker, S. A., & Brownell, A. (1984). Toward a theory of

social support: Closing
conceptual gaps. Journal of Social Issues, 40, 11-36.
Silver, R. c., Wortman, C. 8., & Crofton, C. (1990). The

role of coping in support pro
vision: The self-presentational dilemma of victims of life

crises. In I. G. Sarason,
8. R. Sarason, & G. R. Pierce (Eds.), Social support: An

interactional view (pp. 397
426). New York: Wiley.
Skokan, L. A. (1990). Motivational determinants of social

support provision.
Unpublished doctoral dissertation. University of

California: Los Angeles.
Weiner, B. (1985). An attributional theory of achievement

motivation and emotion.
Weiner, B. (1986). An attributional theory of motivation

and emotion. Berlin: Springer.
Weiner, B., Perry, R. P., & Magnusson, l. (1988). An

attributional analysis of reactions
to stigmas. Journal of Personality and Social Psychology,

55(5) 738-748. This page intentionally left blank
II Self-Efficacy and Human Development
Abramson, L. Y., Seligman, M. E. P., & Teasdale, 1. D.

(1978). Learned helplessness
in humans: Critique and reformulation. Journal of Abnormal

Baltes, M. M., & Baltes, P. B. (Eds.) (1986). Aging and

the psychology of control.
Bandura, A. (1977). Self-efficacy: Toward a unified theory

of behavioral change.
Chapman, M., & Skinner, E. A. (\989). Children's agency

beliefs, cognitive perform
ance and conceptions of effort and ability: Interaction of

individual and develop
mental differences. Child Development, 60, 1229-1238.
Chapman, M., Skinner, E. A., & Baltes, P. B. (1990).

Interpreting correlations between
children's perceived control and cognitive performance:

Control, agency, or means
ends beliefs? Developmental Psychology, 23,246-253.
Connell,1. P. (1985). A new multidimensional measure of

children's perceptions of
control. Child Development, 56, 1018-1041.
Connell, J. P., & Wellborn, 1. G. (1990). Competence,

autonomy, and relatedness: A
motivational analysis of self-system processes. In M.

Gunnar & L. A. Sroufe (Eds.),
Minnesota symposium on child psychology (Vol. 22, pp.

43-77). Minneapolis:
University of Minnesota Press.

Crandall, V. C., Katkovsky, W., & Crandall, V. J. (1965).
Children's beliefs in their
control of reinforcement in intellectual academic

achievement behaviors. Child
Development, 36,91-109.
Crandall, V. C., & Linn, P. L. (1989, April). Children's

achievement orientation: Sex
differences, developmental trends, and emergence of

motivational functions. In J. G.
Nicholls (Chair), Achievement strivings, expectations, and

values in the toddler to
first grade years. Symposium at the Biennial Meetings of

the Society for Research in
Child Development, Kansas City, MO.
Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation

and self-determination in
human behavior. New York: Plenum.
Dweck, C. S., & Wonman, C. B. (1982). Learned

helplessness, anxiety, and achieve
ment motivation: Neglected parallels in cognitive,

affective, and coping responses. In
H. W. Krohne & L. Laux (Eds.), Achievement, stress, and

anxiety (pp. 93-125).
Fincham, F., Hokoda, A., & Sanders, R., Jr. (1989).

Learned helplessness, test anxiety,
and academic achievement: A longitudinal analysis. Child

Findley, M. J., & Cooper, H. M. (1983). Locus of control

and academic achievement:
A literature review. Journal of Personality and Social

Garmezy, N., & Rutter, M. (Eds.) (1983). Stress, coping,
and development in children.
New York: McGraw-Hill.
Gurin, P., Gurin, G., & Morrison, B. M. (1978). Personal

and ideological aspects of
internal-external locus of control. Social Psychology,

41,275-296.
Gunnar, M. R. (1980). Contingent stimulation: A review of

its role in early develop
ment. In S. Levine & H. Ursin (Eds.), Coping and health

(pp. 101-119). New York:
Plenum.
Miller, A. (1985). A developmental study of the cognitive

basis of performance
impairment after failure. Journal of Personality and

Social Psychology, 49,529-538.
Nicholls, J. G., & Miller, A. T. (1984). Development and

its discontents: The differen
tiation of the concept of ability. In J. G. Nicholls

(Ed.), The development of
achievement motivation (pp. 185-218). Greenwich, CT: JAI

Press.
Nolen-Hoeksema, S., Girgus, J. S., & Seligman, M. E. P.

(1986). Learned helplessness
in children: A longitudinal study of depression,

achievement, and explanatory style.
Journal of Personality and Social Psychology, 52,435-442.
Rholes, W. S., Blackwell, J., Jordan, C., & WaIters, C.

(1980). A developmental study
of learned helplessness. Developmental Psychology, 16,

616-624.
Seligman, M. E. P. (1975). Helplessness: On depression,

development, and death. San
Francisco: Freeman Press.
Skinner, E. A. (1985). Action, control judgments, and the

structure of control
experience. Psychological Review, 92, 39-58.
Skinner, E. A. (1990a). Age differences in the structure

of perceived control during
middle childhood: Implications for developmental

conceptualizations and research.
Child Development, 61, 1882-1890.
Skinner, E. A. (1990b). Development and perceived control:

A dynamic model of ac
tion in context. In M. Gunnar & L. A. Sroufe (Eds.),

Minnesota symposium on child
psychology, Vol. XXII (pp. 167-216). Minneapolis:

University of Minnesota Press.
Skinner, E. A., Chapman, M., & Baltes, P. B. (1988a).

Beliefs about control, means
ends, and agency: A new conceptualization and its

measurement during childhood.
Skinner, E. A., Chapman, M., & Baltes, P. B. (l988b).

Children's beliefs about control,
means-ends, and agency: Developmental differences during

middle childhood. Inter
national Journal of Behavioral Development, II, 369-388.
Skinner, E. A., & Connell, J. P. (1986). Development and

the understanding of control.
In M. M. Baltes & P. B. Baltes (Eds.), Aging and the

psychology of control (pp. 35
Skinner, E. A., & Wellborn, J. G. (1991, April). Coping

and perceived control. Paper
presented at the Biennial Meetings of the Society for

Research in Child Development,
Seattle, W A.
Skinner, E. A., & Wellborn J. G. (in press). Coping across

the life span: A motiva
tional perspective. In M. Perlmutter, D. Featherman, & R.

Lerner (Eds.), Life-span
development and behavior, Vol. 12. New York: Academic

Press.
Skinner, E. A., Wellborn, 1. G., & Connell, J. P. (1990).

What it takes to do well in
school and whether I've got it: The role of perceived

control in children's
engagement and school achievement. Journal of Educational

Watson, J. S. (1979). Perception of contingency as a

determinant of social responsive
ness. In E. B. Truman (Ed.), Origins of the infant's

social responsiveness (pp. 33
Weinberg, J., & Levine, S. (1980). Psychobiology of coping

in animals: The effects of
unpredictability. In S. Levine & H. Ursin (Eds.), Coping

and health. New York:
Plenum.
Weiner, B. (1979). A theory of motivation for some

classroom experiences. Journal of
Educational Psychology, 71, 3-25.
Weisz, 1. R., & Stipek, D. 1. (1982). Competence,

contingency, and the development of
perceived control. Human Development, 25, 250-281.

Wellborn, 1. G., & Connell, J. P. (1986). Manual for the
Rochester Assessment
Package for Schools. University of Rochester, Rochester,

NY.
Wellborn, 1. G., Connell, J. P., & Skinner, E. A. (1989).

The Student Perceptions of
Control Questionnaire (SPOCQ). Technical Report, University

of Rochester,
Rochester, NY.
White, R. W. (1959). Motivation reconsidered: The concept

of competence.
Psychological Review, 66, 297-333. Author Notes The

contributions of collaborators are gratefully acknowledged,
both those in
Berlin, Paul Baltes and Michael Chapman; and in Rochester,

James Connell,
Edward Deci, Richard Ryan, and James Wellborn. This paper

was completed
with the generous support of the Max Planck Institute for

Human Development
and Education and a Faculty Scholar's Grant from the

William T. Grant
Foundation. APPENDIX I
SAMPLE ITEMS FROM THE STUDENT PERCEPTIONS OF CONTROL

QUESTIONNAIRE (SPOCQ, Wellborn, Connell, & Skinner, 1989)
Control Beliefs If I decide to Jearn something hard, I can.

(+) I can't get good grades, no matter what I do. (-)
Strategy Beliefs Effort: Ability: Powerful: Others:

Luck: Unknown:
Capacity Beliefs Effort: Ability: Powerful: Others:

Luck: For me to do well in school, all I have to do is
work hard. If I get bad grades, it's because I don't work
hard enough. I have to be smart to get good grades in
school. If I'm not smart, I can't get good grades. If I
want to get good grades, I have to get along with my
teacher. If my teacher doesn't like me, I won't do well in
that class. For me, getting good grades is a matter of
luck. If I get bad grades, it's because I'm unlucky. When
I do well in school, I usually can't figure out why. I
don't know how to keep myselffrom getting bad grades. When
I'm in class, I can work hard. (+) 1 can't seem to try
very hard in school. (-) I think I'm pretty smart in
school. (+) I don't have the brains to do well in school.
(-) I can get my teacher to like me. (+) I just can't
get along with my teacher. (-) I am lucky in school. (+)
When it comes to grades, I'm unlucky. (-) APPENDIX II
SAMPLE ITEMS AND PSYCHOMETRIC PROPERTIES FOR THE COPING

CATEGORIES (From Skinner & Wellborn, 1991)
When something bad happens to me in school (like not doing

well on a test or not
being able to answer an important question in class) ...
Strategizing (5 items, alpha = .68):
Perseverance (5 items, alpha = .73):
Avoidance (5 items, alpha = .62):
Delegation (5 items, alpha = .63):
Projection (5 items, alpha = .66):
Confusion (5 items, alpha = .60):
Impulsivity (6 items, alpha = .68): I try to see what I

did wrong. I slow down and think carefully. I go over
the problem again and again. I can't go on until I've
solved it. I try not to think about it. I put it out of
my mind. I get the teacher to solve the problem. I want
the teacher to tell me the answer. I get real mad at
other people. I say the teacher didn't cover the things
on the test. my mind goes blank. I get all confused. I
just say the frrst thing that comes into my head. I just
do anything. ADULTS' EXPECTANCIES ABOUT DEVELOPMENT AND
ITS CONTROLLABILITY: ENHANCING SELF-EFFICACY BY SOCIAL
COMPARISON lutta Heckhausen The human life span with its
radical changes and structured constraints in control
potential confronts the individual with a highly taxing
coping task. If developmental tasks and everyday
challenges are to be tackled successfully, not only at one
point in the life span, but across shifting developmental
ecologies, the individual has to maintain and balance two
requirements. On the one hand, expectations, plans, and
evaluations have to reflect reality, so as to allow
appropriate action planning. On the other hand, conceptions
about life-span change need to be biased so as to protect
self-efficacy, and thus maintain motivational
prerequisites of action. One way to balance both these
apparently contradictory requirements is strategic social
comparison. Social comparison in terms of age-graded
phenomena is based on normative conceptions about life-span
development. Such normative conceptions help generate
age-graded social reference groups, which provide both, a
realistically tailored and a self-enhancing framework for
personal evaluations and aspirations. Recent research is
presented, which demonstrates the structure and consensual
nature of normative conceptions about adult development.
Moreover, development-related expectations for the self
entail substantial congruence with those ascribed to
"most other people," thus exemplifying a focus on
validity, rather than on excessive self-serving illusion.
However, differences between normative and self-related
expectations reflect a self-enhancement strategy,
particularly with regard to old age, which is regarded a
more threatening period of the life span.
The present chapter addresses adults' expectancies about

age trajectories and
controllability of adult development. Conceptions about

one's own personal
development are contrasted with conceptions about "most

other people's" devel
opment. Accordingly, beliefs about self-efficacy (i.e.,

controllability for self) are
juxtaposed to beliefs about general controllability (Le.,

for "most other people")
of developmental processes. Distinct patterns of self/other

differences in
development-related conceptions reflect different

strategies of using social
comparison to enhance feelings of self-efficacy in either

past, current, or future
personal development (J. Heckhausen & Krueger, 1991).

Self-efficacy, in tum,
is a fundamental requirement for the potential to act and

exert control over one's
environment and self (Bandura, 1982). The human life

course can be conceived as a context for action, action
direct
ed at development (1. Heckhausen, 1990a). Across the life

course, an individ
ual's potential to actively control her environment

undergoes dramatic changes
(Baltes & Baltes, 1986; Baltes, 1987). At birth and during

early infancy the indi
vidual is almost completely unable to manipulate and

change objects in the
external world. Control during this time of life is, to the

most part, exerted by
getting more mature others such as caregivers, to act (1.

Heckhausen, 1987;
Kaye, 1982). During childhood and adolescence the

individual rapidly gains
motor abilities, skiIIs, knowledge, and understanding, and

thereby becomes
increasingly aware of his own competence (Harter, 1975; J.

Heckhausen, 1988)
and autonomy from caregivers and other socialization agents

(e.g., Steinberg &
Silverberg, 1986). In early adulthood it is not only

increasing motorphysical and
cognitive abilities but also the age-graded allocation of

multiple social roles in
the work and family domain (Marini, 1984) which increase

the individual's scope
for exerting control over various life ecologies. During

middle adulthood control
potentials in many domains of life reach their maximum, but
thereby also render
over-ambitious life goals out of reach due to a lack of

remaining life time (Brim,
1988; Neugarten, 1968). Finally, during old age physical

powers decline, motor
and mental activites become slowed (Salthouse, 1985), and

social roles allocated
by society diminish (Riley & Riley, 1986). If

self-efficacy beliefs would simply reflect absolute levels
of personal con
trol, such radical shifts in control potential as

experienced throughout the human
life course would render self-efficacy a very unstable

feature of the human mind.
Self-efficacy, however, has been shown to be a key resource

for motivated action
in humans (Bandura, 1977, 1982, 1986). Self-efficacy can

only serve this
important function if it is a stable and reliable resource,

that is not jeopardized by
obstacles and drawbacks, but instead helps to overcome

difficult spells in life.
The present paper attempts to line out and demonstrate some

of the ways in
which the human mind uses secondary, as opposed to more

direct primary, con
trol strategies (Rothbaum, Weisz, & Snyder, 1982) in order

to maintain long-term
self-efficacy in spite of constrained and declining control

potential over the life
course (J. Heckhausen & Schulz, 1991; Schulz, Heckhausen,

& Locher, in press).
Special emphasis is given to strategies of secondary
control relying on selective
social comparison. These issues are discussed with regard

to the role of lay
persons' normative conceptions about development in

adulthood. The life course provides a unique time-graded
structure of opportunities and
risks for developmental gains and losses. Attempts to

actively influence one's
own development are constrained in specific age-related

ways. Some of these
restrictions are fairly rigid, such as the timing of school

entry or retirement, some
are more lenient. In any event, these age-related

constraints provide guideposts
for assessing an individual's current developmental status

in the life course, and
for advising future life planning. Three types of life

course constraints can be
discerned: (I) constraints in terms of the life time

remaining till death (absolute),
(2) constraints associated with chronological age

(cross-sectional or vertical), and
(3) constraints resulting from sequential patterns

(longitudinal or horizontal)
(J. Heckhausen, 1990a; J. Heckhausen & Schulz, 1991).

First, life time
remaining till death restricts the potential future time

extension of developmental
goals and life plans. This is relevant, for instance, when

in midlife the feasabiIity
of career changes is considered. Second, age norms set

deadlines for various life
events, and thereby slice the life span in vertical

segments. Such norms can be
set on a societal level, such as in the case of school

entry or retirement (Hagestad
& Neugarten, 1985; Neugarten, Moore, & Lowe, 1965), or be

the product of uni
versal biological processes such as in the case of puberty

as the onset and meno
pause as the close of fertility. Third, age-sequential

(horizontal) constraints can
also be identified on the societal and the biological

plane. Biologically they are
shown in universal sequential patterns of maturation and

aging. On the societal
plane, age-sequential constraints result from the

canalization of development and
life course patterns into developmental or biographical

tracks, involving more or
less fixed sequences and timing of life events (Geulen,

1981; Mayer, 1986), for
instance during the transition to adulthood (Marini, 1984;

Modell, Furstenberg, &
Hershberg, 1976). These constraints in and of

themselves-not even considering the substantial
losses associated with aging-reveal that we are far from

being in complete con
trol over our lives. Instead, we have to adapt our hopes,

goals, and plans to a
fairly restricted repertoire of options over which we have

at least some control.
If our feelings of self-efficacy were perfect reflections

of all these constraints we
would be close to paralytic helplessness. In order to

generate and maintain the
potential for action directed at development and the life
course, two requirements
have to be fulfilled. On the one hand, the individual

needs a fairly valid concep
tion about options, possible life paths, suitable action

means, and probability of
success, in order to generate effective action. On the

other hand, the motivational
management of action requires feelings of self-efficacy and

thus, hope for suc
cess; otherwise the individual would be overwhelmed by

uncertainty and fear of
failure. Given the constraints and radical life-span shifts

in control potential,
these two requirements are bound to get into conflict (J.

Heckhausen & Schulz,
1991). When, for instance, physical strength drastically

declines in old age, per
sonal aspirations for athletic activities have to reflect

this loss in order to keep in
touch with reality. However, feelings of self-efficacy and

self-worth also have to
be protected in order to motivate the individual to keep up

appropriate activity. One most effective way in which both
these partly contradictory require
ments can be met is strategic social comparison. Festinger

(1954) has proposed
that social comparison may serve self-assessment and

self-improvement func
tions. In order to assist self-assessment the individual

compares with others who
are similar on relevant performance dimensions, so that the

evaluation of one's
own performance is calibrated. Self-improvement is

promoted if one compares
with others who are a little superior on a relevant

dimension. Incidently, such
upward social comparison converges with the selection of a

just-above-medium
level of aspiration, which is identified as most promotive

for performance and
learning by achievement motivation research (see review in

H. Heckhausen,
1991). Wood (1989) has recently proposed a three-fold model

of social compari
son adding self-enhancement as a major function of

comparison to self
assessment and self-improvement, as proposed by Festinger

(1954). Self
enhancement via social comparison is achieved by selecting

a reference group
which is inferior to oneself on relevant dimensions.

Vis-a-vis such a reference
group one can see oneself in a more favorable light, even

if one has experienced
a loss of self-esteem. In fact, recent empirical research

has shown that such
"downward comparisons" (Wills, 1981) are particularly

sought out under condi
tions of threatened life or self-esteem, and impaired

health. Severe disability
(Schulz & Decker, 1985), life-threatening illnesses (Taylor

& Lobel, 1989;
Taylor, Wood, & Lichtman, 1984), or crime-related

victimization (Burgess &
Holmstrom, 1979) represent such eliciting conditions for

"downward
comparisons. " With regard to age-graded phenomena
strategic social comparison is espe
cially suitable. Throughout their life course individuals

experience major intra
individual changes, for better and worse. Many of these

changes are universal
and, thus, shared with age peers. Therefore, the peer group
provides a most suit
able reference for self-assessment. If one compares

oneself with age peers the standards of aspiration become
realistic. At the same time feelings of self-efficacy are

protected, because one
compares oneself with others who undergo similar

age-related changes (Baltes, in
press; Baltes & Baltes, 1990). The self-enhancement

function of social compari
son becomes especially salient in old age, when

self-esteem is threatened by
aging-related decline. Selecting a suitable comparison

group for downward
comparison in old age, however, is hardly a problem, since

old age and aging is
viewed in a negative stereotype manner. Survey research

about the image of old
age and aging has revealed that old people view old age as
a bleak period of the
life span for "most old people," yet see themselves as

favorable exeptions (Harris
& Associates, 1975, 1981). O'Gorman (1980) has extended

this conclusion by
showing that elderly estimated "most other old people's"

problems in various life
domains (e.g., health, finances, loneliness) higher, if

they themselves reported
experiencing the respective problem. Thus, self-enhancement
via downward
social comparison may be a common and effective way of

coping with aging
related decline (J. Heckhausen & Krueger, 1991).

Generalized images of age groups are needed to generate
mental representa
tions about a social comparions group. In this context,

normative conceptions
about development playa crucial role, because they provide

common reference
frames for expectations and evaluations of developmental

change. One of the
unique features of the life course as a context for action

is that, although the III
developmental tasks involved might not be currently

relevant to anyone of us,
they will be or have been at some point in our life

course. The life course and its
various challenges are therefore personally relevant to

everyone. Some of its per
iods, such as old age, might seem somewhat far away for a
young person, for
instance. However, they will eventually become the critical

context for that per
son's own life management. Expectations for oneself as well

as social compari
son with regard to age-graded phenomena is based on

normative conceptions
about development in adulthood. The human life course

provides a more or less
common repertoire of sequential developmental tasks to the

members of a given
society (J. Heckhausen, I 990b ). Therefore, it is
essential both, for the individual
and for the society at large to have conceptions about

development and the life
course which are commonly shared. Such normative

conceptions about develop
ment and the life course represent "social constructions

of reality" (Berger &
Luckmann, 1966). They provide the reference frame for the

developmental
changes in oneself and for social comparisons both within

and across age groups. Normative conceptions about
life-span development probably influence our
thinking and actions with respect to all three planes in

time, future, present, and
past. With regard to the future, they set the stage for
one's own developmental
prospects. Certain developmental changes, such as learning

a language, might
appear more obtainable at some ages than at others. The

individual will therefore
try to time respective efforts accordingly, or if that is

unfeasable, will know that
special effort has to be invested to bring about the

untimely change. Moreover, if
certain changes, such as becoming forgetful, are expected

to occur inevitably at a
certain age, the individual will not expect to modify his

or her developmental tra
jectory and thus will not invest effort. This way,

normative beliefs about develop
ment form our expectations about what we can do, what we

cannot do, and where
to invest special effort in trying to take charge of our

developmental future. Moreover, normative beliefs about
development could serve as a frame of
reference for evaluating current developmental change in

the self. Think, for in
stance, of undesired developmental changes, such as

becoming forgetful in
advanced age, or being worn out by mid-life crisis. Here,

normative expectations
might lead to a downgrading of one's peer reference group,

and thereby may help
us to maintain self-esteem and well-being in spite of

unpleasant developmental
experiences. If we believe that most of our peers are

experiencing the same, we
are less likely to blame ourselves; and if we suffer less

decline than, we think,
most of our age peers do, we actually could pride

ourselves to have come up
against the stream of aging. Finally, normative beliefs

about development might help us to reconstruct
and interpret our biographical past. The sequence of

biographical events we
encountered in our lives appears as one path in a limited

repertoire of life courses
possible or likely in a given socio-cultural context

(Kohli, 1981). This way,
normative conceptions of the life course provide structure

to individuals' biogra
phies. They inform us, in what ways we are the same, and in
which ways we are
different from most other people. Both aspects are critical

for our personal well
being. We want to be connected to the social world around

us, but still we also
want to be unique and somehow different from others
(Campbell, 1986; Snyder
& Fromkin, 1980). Thus, the optimal relation between what

is perceived to be
normative or typical and what is descriptive of the self

probably is one which
strikes the balance between consensus and uniqueness. The

examples just given about the role of normative
conceptions about
development for future life planning, current life

management, and past life
review, illustrate two types of phenomena. On the one

hand, normative devel
opmental conceptions are used to provide valid information

about chances and
risks, so that the individual could optimize his or her

active developmental inter
ventions. On the other hand, normative conceptions serve to

console the person
by favorable comparison with less fortunate others, or by

fending off blame.
These two modes of exerting control exemplify the

distinction between primary
and secondary control (Rothbaum et aI., 1982) for the

context of development
related conceptions and action. Primary control refers to

direct action on the
external world. Conversely, secondary control refers to

cognitive adjustments in
terms of mental representations. The target of primary

control is the external
world, whereas the target of secondary control is the own

self, one's goals, well
being, interpretations, and most importantly one's
self-efficacy. This distinction
between primary and secondary control is related to

Piaget's conceptual distinc
tion between assimilative and accomodative processing

(Piaget, 1985). Brandt
stadter and Renner have used these Piagetian concepts to

characterize primary
and secondary modes of coping in middle-aged adults, and

identified comple
mentary age trends in a cross-sectional study covering

middle adulthood
(Brandtstadter & Renner, 1990; see also Brandtstadter, this

volume). Given the radical life-span changes in the
potential to control one's environ
ment and development, it would appear functional if there

was a gradual shift
from a preference for primary control strategies to a

predominance of secondary
control strategies at later periods of the life course (J.

Heckhausen & Schulz,
1991; Schulz et aI., in press). This hypothesis converges

with Brandtstadter and
Renner's finding of decreased importance of assimilative

as compared to accom
odative control strategies throughout middle adulthood

(Brandtstadter & Renner,
1990). To be sure, at each period of the life span both

strategies are necessary to
cope with the challenges and constraints characteristic for

that life-span segment.
Moreover, primary and secondary strategies of control have

to be balanced such
that action is optimized, while self-efficacy is protected.

Leaning too much into
either direction bears major hazards to developmental

potential. Being too realis
tic about the constraints in control exposes the individual

to depression (Alloy &
Abramson, 1979; Lewinsohn, Mischel, Chaplin, & Barton,

1980; Taylor, 1989),
and thus reduces motivational resources for action.

Conversely, if self-efficacy is
boosted so much that the individual loses contact to

reality, cognitive require
ments for planning and execution of action are

jeopardized. Some recent empirical work at the Max Planck
Institute will be reported to
demonstrate, how in the context of adult development

primary and secondary
control strategies are balanced in a functional way, such

as to optimize long-tenn
action potential in tenns of both, its cognitive and

motivational requirements. In
this research, primary and secondary control strategies are

identified in social
comparison as operationalized in differences between

development-related
expectations for the self as compared to "most other

people." We can expect to find both strategies of
successful life management reflected
in people's conceptions about development, and particularly

in the way they view
their own developmental status vis-a-vis the nonnative

context of developmental
expectations for "most other people." Specifically, one can

expect primary con
trol to be reflected in a focus on the veridicality of

development-related concep
tions. An effort to arrive at veridical conceptions about

development would be
shown in largely similar conceptions for the self and "most

other people." Thus,
developmental expectations for the self would not be

unrealistically biased to
wards optimism. Expected developmental trajectories for

self and others should
be largely congruent. And perceived controllability for

late life change should be
lower than for early onset developmental change. I

However, conceptions about development should, to some
degree, also
reflect secondary control processes. First, an overall

optimistic view of devel
opmental prospects in adulthood can be expected. This would

be shown in a pre
dominance of expected desirable as compared to undesirable

developmental
changes. Second, with regard to the comparison of self and

other, more devel
opmental gains should be expected for the self than for

"most other people." And
third, especially with regard to old age, when expected

developmental losses
might prevail, self-related expectations should be more

optimistic than expecta
tions held for "most other people." Before dealing with

these specific findings
the first question in our research is of course: Are there

at all common normative
conceptions about development in adulthood that most

members of a given
society would agree on? And if yes, what do these
conceptions entaiJ? Three studies were conducted. The
first two dealt with adults' and adoles
cents' expectations about developmental gains and losses

in adulthood (J. Heck
hausen, I 990c; J. Heckhausen, Dixon, & Baltes, 1989; J.

Heckhausen, Krueger,
& Hosenfeld, 1989; Hosenfeld, 1988). The third one focused

on the perceived
controllability of such expected changes (J. Heckhausen &

Baltes, in press). The procedures and materials used in
these studies were similar. The sample
contained young adolescents between II and 17 years of age

(Hosenfeld, 1988),
young adults between 20 and 35 years of age, middle-aged

adults between 40 and
55 years, and old adults above 65 years (J. Heckhausen, I

990c; 1. Heckhausen &
I It should be noted at this point that we have no means of

knowing to what degree con
ceptions about development are veridical (i.e., mirror

reality), and whether optimistic or
self-enhancing views of development are true or false.

Life-span developmental psy
chology is far from being able to chart all reI event

developmental trajectories; and most
likely this will never be possible nor make sense.

Moreover, all relevant assessments
biases (Krueger & Heckhausen, 1991).
Baltes, in press; J. Heckhausen, Dixon, & Baltes, 1989).

Subjects were given a
large list of adjectives denoting a variety of

psychological attributes, for instance,
intelligent, wise, forgetful, bitter, and mentally healthy.
The subjects were asked to rate each of these attributes
with regard to the
following questions: (I) How much does this attribute

increase in strength across
the adult life span, or does it not increase? (2) How

desirable versus undesirable
is a developmental increase in this attribute? (3) At which

age does this develop
mental increase typically begin, and at which age does it

end? And finally, (4) to
what extent can one control, that is promote or hinder, the

developmental
increase in this attribute 2 ? We had expected to find a

pattern of life-span expectations which would
reflect a negative view of aging, that is, a target-age

related shift from a predomi
nance of gains to a predominance of losses. However, the

image of old age
should not be entirely bleak, so that the elderly would not

be left to despair.
Figure I shows the overall ratio of gains and losses across

the life span. Note
that the abscissa shows the target ages, not the subjects'
ages. IIlU 90 80 70 60 50 40 30 20 10 20 .10
40 50 60 70 80 90 Age
Figure 1 Expected gains and losses across the adult life

span (from: J. Heckhausen, Dixon, & Baltes, 1989). As
shown in Figure I, there is, as expected, a gradual shift
from a predomi
nance of gains expected for early adulthood to a

predominance of losses in old
age. Thus, people's conceptions about adults' developmental

prospects reflect
aging-related decline. However, Figure I also shows

something else. Overall,
the expected gains greatly outnumber the losses. Even in

the 80s, about 20% of
2 Perceived controllability was not assessed in the study

on adolescents (Hosenfeld,
1988). Gains Losses 10 P e r c e n t a g e s o f G a i n

s a n d L o s s e s
the total changes are expected to be gains. Thus, although

the general pattern of
normative beliefs about adult development reflects a

negative view of aging, peo
ple of all age groups see some potential for growth even
in advanced age. What
about the perceived controllability of these expected

changes? Controllability
would appear to be the key variable to ameliorate the

pretty unfavorable view of
developmental prospects in old age.
Figure 2 Three-dimensional representation of mean ratings

of perceived controllability, desirability, and expected
onset age for 163 change-sensitive attributes (from: 1.
Heckhausen & Baltes, in press). Figure 2 provides a
three-dimensional display of each attributes' mean
ratings on three variables: Expected age of change onset,

perceived controllabil
ity, and desirability. Each of the arrows refers to one

attribute. The onset ages 20
through 60 years are shown on the horizontal scale in

front; desirability from low
(very undesirable = 1) to high (very desirable = 9) at the

side; and controllability
on the vertical scale, ranging from somewhat below medium

(medium controlla
bility = 5) to high (very controllable = 9). The

controllability scale itself, which
starts at a value of 4, shows that no attributes were

perceived to be really low in
controllability. Most mean ratings in fact were above the

level of "medium con
trollability." This speaks to a fairly optimistic view

about the scope of controlla
bility, and thus plasticity of developmental change

overall, even for advanced age. Second, developmental
changes expected to occur later in life were expected
to be less desirable: As shown in Figure 2, arrows at the

later target ages origi
nate in front segments which correspond to low levels of

desirability. This illus
trates the shifting ratio of gains and losses, also

depicted in Figure I. 4 9 8 7 6 5 4 3 2 1 20 30 40 50 60
Onset Age 7 6 5 C o
n t
r o
l l a
b i
l i t
y 7 D e s i r a b i l i t y F i g u r e 3 M e a n r a t i n
g s o f p e r c e i v e d c o n t r o l l a b i l i t y a n
d d e s i r a b i l i t y f o r 1 6 3 c h a n g e s e n s i
t i v e a t t r i b u t e s ( f r o m : J . H e c k h a u s
e n & B a l t e s , i n p r e s s ) . h i g h r = . 6 4 7 6
Controllability m e d i u m b o s s y s t i l t e d m o r a
l i s t i c b i t t e r a b s e n t m i n d e d f o r g e t
f u l c o m p l i c a t e d q u e e r d i s t u r b e d 4 v
e r y u n d e s i r a b l e 2 3 4 5 D e s i r a b i l i t y
s e n s i t i v e 6 7 8 v e r y d e s i r a b l e i n t e l
l i g e n t m e n t a l l y h e a l t h y w i s e i n t e r
e s t s w i d e e d u c a t e d o r g a n i z e d t h r i f
t y w e l l r e a d k n o w l e d g e a b l e With regard
to perceived controllability there is also an age-related
shift. At
increasing target ages the arrows become shorter and
shorter, indicating lower
degrees of perceived controllability. Thus, later age

periods in the adult life span
are associated with less desirable and less controllable

developmental changes.
Note however, that even for changes expected very late in

life a substantial de
gree of control was still expected, since even the short

arrows indicate about
medium controllability. In order to illustrate more

specifically the relationship between the perceived
controllability and the desirability of developmental

changes, Figure 3 provides a
two-dimensional plot showing mean ratings of

controllability and desirability for
each attribute. In Figure 3 some of the attributes are

identified. The undesirable
and less controllable ones in the lower left corner, such

as disturbed, absent
minded, and forgetful, are prototypical features of

decline in old age. Interest
ingly enough, the opposite cluster, that is those

attributes which are very
desirable and appear most controllable (see upper right

comer), are virtues of
life-long further education: well-read, knowledgeable,

educated, wide interests;
but not wisdom, which is highly desirable but appears by

far not as controllable. Figure 3 also illustrates the
strong positive relation between desirability and
perceived controllability of attributes. Desirable

psychological features appeared
clearly more controllable than undesirable features. This
is reminiscent of the
well known "attributional bias": People like to take credit

for the positive events,
but do not want to be blamed for negative ones (Bradley,

1978; H. Heckhausen,
1987; Kelley & Michela, 1980; Snyder, Stephan, &

Rosenfield, 1978; Zucker
man, 1979). Such an attributionaI pattern is, of course,

conducive to successful
life management. It encourages the striving for

improvement, but also consoles
and avoids self-blame at times of losses. Thus, it is

characteristic of the dual
function of development-related conceptions: information

and consolation, pri
mary and secondary control. Interestingly, the relationship

between desirability
and perceived controllability was closer for the old and

the middle-aged when
compared to the young adults (J. Heckhausen & Baltes, in

press). For old adults,
who are surely confronted with more losses than younger

adults, this attributional
bias would be more essential than for younger people. The

age difference in the
degree of the attributional bias demonstrates the

functionality of normative con
ceptions about development in terms of buffering potential

negative effects of
aging on self-efficacy. In sum, on the one hand, normative

conceptions about developmental change
and its controllability do reflect a negative view of

aging. With regard to old age
processes of decline clearly predominate over processes of
growth, and less
controllability is expected for losses, losses that occur

increasingly in old age.
On the other hand, at all age periods some, however

restricted, potential for
growth is expected. And what is more, even the most severe

losses seem not in
evitable, but instead leave some hope for plasticity

subject to active intervention. In a second step in our
research program, we extended our research paradigm
to include self-related conceptions about development. In

a recent study, our
major goal was to explore how expectations about one's own

development are
related to normative conceptions about the development of

"most other people"
(J. Heckhausen & Krueger, 1991; Krueger & Heckhausen,

1991). As mentioned
before, we view normative beliefs about development as

social constructions of
reality (Berger & Luckmann, 1966), which acquire the

function of reference
frames. They can guide action directed at the future, they

help us in evaluating
current developmental change, and they provide the scaffold

to build interpreta
tions of the past. We therefore predict that expected

developmental trajectories
for the self are similar to expected trajectories for

"most other people." However,
in certain respects self-related trajectories should differ

from those attributed to
"most other people." As argued already, normative beliefs
about development
might also serve to foster self-enhancement. For instance,

if the developmental
course of the typical age peer is expected to show

decline, the self might be
viewed as a praisable exception to this normative age

trend. Thus, via social
comparison with the expected typical age-related change,

self-esteem is enhanced
or at least maintained, even when facing actual age-related

decline in functioning. These theoretical expectations
were investigated in a study comparing
development-related conceptions regarding the self with

those ascribed to "most
other people" (J. Heckhausen & Krueger, 1991; Krueger &

Heckhausen, 1991).
The subject sample included 180 subjects with 60 subjects

from each of three
adult age groups: young (age range: 20 to 35 years),

middle-aged (age range: 40
to 55 years), and old adults (age range: above 60 years).

The age groups were
equally divided by gender and by three levels of

educational background: People
who finished high school, people with some lower college

education, and people
who had earned college degrees. For reasons of

parsimoniousness, between-sub
ject differences due to age, sex, or educational background

will not be discussed
here but elsewhere (J. Heckhausen & Krueger, 1991; Krueger

& Heckhausen,
1991). Instead, this article focuses on the within-subject

factors in our design,
that is, the difference between self and other, and the
different changes expected
for various decades of the adult life span. Like in the

first set of studies we used an attribute rating format.
The var
iables, on which the attributes were rated, included "self

description at present,"
"desirability," "perceived controllability" rated

separately for the self and for
"most other people," and the "expected change from 20 to

90 years of age" also
rated separately for the self and "most other people."

Ratings of expected change
could be given as increases, ranging from +1 to +3,

stability given as 0, or de
creases, ranging from -I to -3. Finally, we asked the

subjects to indicate up to ten
"developmental intentions," that is, to select attributes

on which they intended to
strive for a change (see also Hundertmark, 1990). The pool

of attributes was structured by five major psychological
dimen
sions: Extraversion, Agreeableness, Conscientiousness,

Emotional Stability, and
Intellectual Functioning---commonly known as Norman's Big

Five (Norman,
1963). Each of these factors consisted of two scales with

ten attributes each, one
scale for the desirable attributes, and one for the

undesirable attributes. Analyses
of the data showed that our previous finding of a shifting

ratio of expected gains
and losses across the adult life span was replicated. At

increasing target ages, less
and less gains, and more and more losses were expected,
although as in our
previous studies, some potential for growth was envisaged

even for old age. Q,I 01) C C':I .c U "0 ~ Col
<II Q. >( "-l 0.5 <II ~ 0,4 C':I .c U "0 0,3 ~
c.J <II Q. ~ 0,2 0,1 0.0 0.6 0,5 '" 0,4 Q,I '"
C':I <II 0.3 ... Col C 0,2 0,1 fIl 0,0 <II fIl
C':I ·0,1 <II ... Col Q,I Q -0.2 Extra. Agree.
Consc. Emot. Intel!. Dimension Extra. Agree. Co lise.
Emot. Inlel!. Dimension Self Olher Self Other
Figure 4 Expected change for self and "most other people"

on five psychological dimensions, separately for desirable
(upper panel) and undesirable (lower panel) attributes
(from: J. Heckhausen, Krueger, & Hosenfeld, 1989). 0.5
With regard to the comparison between self and "most other
people" in terms
of expected developmental prospects, one should first

consider the net change,
that is, the expected change averaged across age decades

in adulthood. In
accordance with the self-enhancement hypotheses, people

should expect more in
creases for the self as compared to other people in the

case of desirable attributes,
and less increases or more decreases for the self when

undesirable attributes are
concerned. Figure 4 shows developmental expectations for

self and "most other
people" relating to desirable (upper panel) and

undesirable attributes (lower
panel) of the five Nonnan dimensions. Ratings for the self

are shown in the dark
bars; those for "most other people" are given in lighter

bars. On the vertical scale
the expected net increase (i.e., averaged across decades)

is plotted. As one can see in the upper panel of Figure
4, more increase in the desirable
attributes were expected for the self than for "most other
people." However, sep
arate analysis for each dimension revealed that this

difference held statistically
only for the dimensions Extraversion and Intellectual

Functioning. The lower
panel shows the complementary picture for undesirable

attributes. Here, the ver
tical scale represents the expected net increase or

decrease in undesirable attri
butes. For "most other people" more increases in

undesirable attributes were
expected than for the self. And accordingly for the

decreases: Less decreases in
undesirable features were expected for "most other people"

than for the self.
This pattern of finding holds statistically for each

dimension, except for Con
scientiousness. Hence, we find evidence for a

self-enhancement effect, both for
the desirable and for the undesirable attributes. The

developmental prospects for
the self are consistently seen more positively than those

for "most other people." Would this self-enhancement
tendency also be reflected in perceptions of
self-efficacy versus general controllability? Figure 5

gives the mean perceived
controllability of desirable and undesirable attributes for

the self and "most other
people." We see here again a self-enhancing tendency in

perceptions of controlla
bility. More controllability is ascribed to the self as

compared to "most other
people" both, for undesirable and desirable features.
Moreover, our previous
finding of an attributional bias, that is, desirable

attributes appear more control
lable than undesirable attributes, is replicated. The

next question is, how this self-enhancing view is
reflected in the curves
of developmental change across decades, expected for the

self and for "most
other people." We would expect to find a pattern of curves

that reflects a post
ponement of developmental decline, and an extended

maintenance of develop
mental growth for the self than for "most other people."
Figure 6 depicts the
expected change curves for self and "most other people"

separately for desirable
and undesirable attributes. The ordinate indicates the

average degree and direc
tion of change: Expected increases are shown above the

dotted line, and
expected decreases below. The trajectories expected for the

self are plotted in
solid lines. Those expected for "most other people" are

shown in dotted lines.
Figure 5 Perceived controllability of desirable and

undesirable attributes for self and "most other people."
Figure 6 Expected change for self and "most other people"

across the adult life span in desirable and undesirable
attributes (from: 1. Heckhausen. Krueger, & Hosenfeld,
1989). 7 6 5 desirable undesirable Attributes Self Other P
e r c e i v e d C o n t r o l l a b i l i t y desirable
attributes for others for self 2 1 0 -1 D e c r e a s e s I
n c r e a s e s E x p e c t e d C h a n g e undesirable
attributes for self for others 20'S 30's 40's 50's 60's
70's 80's Age Period in Decades 1. Heckhausen At first
glance, one is struck by the high degree of similarity
between the
expected trajectories for the self and "most other people"

.. This suggests that
developmental expectations for the self closely reflect

common and general con
ceptions about what normaHy or typically happens as people

grow older. It thus
seems, that conceptions about development in "most other

people" might indeed
serve as a normative framework for sketching one's own

development. However, there are also some differences. In
fact, the pattern of curves con
firms our prediction both for the desirable and the

undesirable attributes. Starting
from the age period of the 50s, the decline in the

desirable attributes is viewed as
less severe for the self than for "most other people." And
also as early as for the
50s, increases in undesirable attributes are expected to

be greater in "most other
people" than in the self. To summarize our findings: In

a sequence of studies we have shown, that
normative beliefs about development reflect common

knowledge, widely shared
among people varying in age, gender, and social strata.

According to this com
mon knowledge about development, aging throughout adulthood

is a process of
ever increasing risks for losses, and ever decreasing

chances for gains. However,
in spite of this negative view of the aging process, common

conceptions about
life-span development also involve optimism. Optimism with

regard to potential
for growth even in very advanced age, and optimism with

regard to personal
control to counteract decline. Expectations about one's

own developmental prospects largely follow similar
trajectories as those expected to hold for "most other

people." However, one's
own development is viewed more optimistically. We believe

to have more per
sonal control over our own development than we ascribe to

"most other people.
We expect more gains and fewer losses for the self than for
"most other people."
This self-enhancement tendency is particularly salient for

the second half of the
adult life span, when increases in undesirable attributes

are expected to occur
earlier in "most other people" than in the self; likewise

decline in positive
features appears delayed for the self when compared to

"most other people." To conclude, I would like to briefly
recapitulate what was said at the begin
ning of the chapter. The life course has a unique

age-related structure of chal
lenges. It confronts us with ever-changing demands, some

of which we can face
and meet with active attempts to intervene; but others

seem to or actually do
leave little scope for active, assertive control. Instead

it seems they require some
sort of re-interpretation, in order to protect a sense of

self-efficacy, and thus
safeguard motivational prerequisites of successful action.

To successfully master a life course one needs to have two
sides. On the one
hand, the efficient agent, who holds realistic expectations

about his develop
mental prospects and strives for attainable goals. But

also, on the other hand, the
skillful self-manipulator, who always, even when confronted

with losses,
manages to maintain a balance of affect, and keeps up

perceptions of self
efficacy. The task of balancing the two needs throughout

the life course and in
spite of radical changes in actual control potential, may

sometimes appear like
fitting a square peg in a round hole. However, the human

mind finds ingenious
ways to resolve the apparent paradox. One most suitable way

is the strategic use
of social reference groups, such that under conditions of

stress one can make
onself feel better by downward comparison. In the service

of control and agency we need to hold fairly realistic
expecta
tions about potential gains and losses. However, the

constraints and losses
encountered throughout the life span are so manifold that

one cannot afford to be
a realist. It looks like, the trick is to strike the

balance. And that, I believe, the
Alloy, L. B., & Abramson, L. Y. (1979). Judgement of

contingency in depressed and
non-depressed students: Sadder but wiser? Journal of

Experimental Psychology, 108,
441-485.
Baltes, M. M., & Baltes, P. B. (Eds.). (1986). The
psychology of control and aging.
Baltes, P. B. (1987). Theoretical propositions of

life-span developmental psychology:
On the dynamics between growth and decline. Developmental

Psychology, 23,611
626.
Baltes, P. B. (in press). The many faces of human aging:

Toward a psychological
culture of old age. Psychological Medicine.
Baltes, P. B., & Baltes, M. M. (1990). Psychological

perspectives on successful aging:
The model of selective optimization with compensation. In

Baltes & Baltes (Eds.),
Successful aging: Perspectives from the behavioral sciences

(pp. 1-34). New York:
Cambridge University Press.

Bandura, A. (1982). Self efficacy in human agency.

American Psychologist, 37, 122
147.

Berger, P., & Luckmann, T. (1966). The social construction

of reality. New York:
Doubleday.
Bradley, G. W. (1978). Self-serving biases in the

attribution process: A re-examination
of the fact or fiction question. Journal of Personality and

Social Psychology, 36, 56
71.
Brandtstlidter, 1., & Renner, G. (1990). Tenacious goal

persuit and flexible goal adjust
ment: Explication and age-related analysis of assimilative

and accomodative
strategies of coping. Psychology and Aging, 5,58-67.
Brim, O. G. (1988). Losing and winning. Psychology Today,

9,48-52.
Burgess, A. W., & Holmstrom, L. (1979). Rape: Crisis and

recovery. Bowie, MD:
Brady.
Campbell, J. (1986) Similarity and uniqueness: The effects

of attribute type, relevance,
and individual differences in self-esteem and depression.

Festinger, L. (1954). A theory of social comparison

processes. Human Relations, 7,
117-140.
Geulen, D. (1981). Zur Konzeptionalisierung

sozialisationstheoretischer Ent
wicklungsmodelle. In J. Matthes (Ed.), Lebenswelt und

soziale Probleme. Verhand
lungen des 20. Deutschen Soziologentages

IConceptualization of developmental
models in socialization theory] (pp. 537-556). Frankfurt:

Campus.
Hagestad, G. 0., & Neugarten, B. L. (1985). Age and the

life course. In R. H. Binstock
& E. Shanas (Eds.), Handbook of aging and the social
sciences (pp. 35-61). New
York: Van Nostrand Reinhold.
Harris, L., & Associates (1975). The myth and reality of

aging in America. Washington,
DC: National Council on the Aging.
Harris, L., & Associates (1981). Aging in the Eighties:

America in transition.
Washington, DC: National Council on the Aging.
Harter, S. (1975). Developmental differences in the

manifestation of mastery motivation
on problem-solving tasks. Child Development, 46, 370-378.
Heckhausen, H. (1987). Causal attribution patterns for

achievement outcomes: Individ
ual differences, possible types, and their origin. In F.

E. Weinert & R. H. Kluwe,
(Eds.), Metacognition, motivation, and understanding (pp.

143-184). Hillsdale, N.J.:
Erlbaum.
Heckhausen, H. (1991). Motivation and action. New York:

Springer.
Heckhausen, 1. (1987). Balancing for weaknesses and

challenging developmental
potential: A longitudinal study of mother-infant dyads in

apprenticeship interactions.
Developmental Psychology, 23, 762-770.
Heckhausen, J. (1988). Becoming aware of one's competence

in the second year:
Developmental progression within the mother-child dyad.

International Journal of
Behavioral Development, I I, 305-326.

Heckhausen, 1. (1990a). Life course patterns, developmental
projects. and control
beliefs in Eastand West-Berliners born between 1920 and

1970: A study proposal.
Unpublished grant proposal, Max Planck Institute for Human

Development and
Education, Berlin.
Heckhausen, 1. (I 990b). Erwerb und Funktion normativer

Vorstellungen tiber den
Lebenslauf: Ein entwicklungspsychologischer Beitrag zur

sozio-psychischen Kon
struktion von Biographien IAcquisition and function of

normative conceptions about
the life course: A developmental psychology approach to the

psychosocial construc
tion of biographies]. Kolner Zeitschrift fur Soziologie

und Sozialpsychologie, Sonder
heft 31: (K.-U. Mayer fEd.], Lebensverlaufe und sozialer

Wandel) (pp. 351-373).
Opladen: Westdeutscher Verlag.
Heckhausen, 1. (l990c). Entwicklung im Erwachsenenalter

aus der Sicht junger, mittel
alter und alter Erwachsener fDevelopment in adulthood as

perceived by young,
middle-aged, and old adults]. Zeitschriftfur

Entwicklungspsychologie und Piidagogi
scher Psychologie, 22, 1-21.
Heckhausen, 1., & Baltes, P. B. (in press). Perceived

controllability of expected psy
chological change across adulthood and old age. Journal of

Gerontology:
Psychological Sciences.
Heckhausen, 1., Dixon, R. A., & Baltes, P. B. (1989). Gains
and losses in development
throughout adulthood as perceived by different adult age

groups. Developmental
Psychology, 25(1), 109-121.
Heckhausen, 1., & Krueger, J. (1991). Similarity and

differences in developmental
expectations for the self and "most other people":

Age-grading in three functions of
social comparison. Manuscript submitted for publication,

Max Planck Institute for
Human Development and Education, Berlin, FRO.
Heckhausen, 1., Krueger, J., & Hosenfeld, B. (1989, July).

Normative beliefs about
development: Their relation to developmental expectations

for the self. Paper pre
sented at the Tenth Biennial Meeting of the Society for

the Study of Behavioral
Development (ISSBD) in Jyvaskyla, Finland.
Heckhausen, J., & Schulz, R. (1991). Primary and secondary

control: Shifting trade
offs across the life span. Unpublished Manuscript, Max

Planck Institute for Human
Development and Education, Berlin, FRO.
Hosenfeld, B. (1988). Personlichkeitsveriinderungen im

Erwachsenenalter aus der Sicht
Jugendlicher IPersonality change in adulthood as perceived

by adolescents].
Unpublished Masters thesis, Freie Universitat Berlin.
Hundertmark, 1. (1990). Entwicklungsbezogene /ntentionen

im Lebenslauf' Selbstbild
und normative Entwicklungsvorstellungen als
EinflujJfaktoren [Development-related
intentions throughout the life span: The impact of

self-concept and normative con
ceptions about development]. Unpublished Master's thesis,

Max Planck Institute for
Human Development and Education, Berlin.
Kaye, K. (1982). The mental and social life of babies: How

parents create persons.
Chicago: The Harvester Press.
Kelley, H. H., & Michela, J. L. (1980). Attribution theory

and research. Annual Review
of Psychology, 31,457-501.
Kohli, M. (1981). Zur Theorie der biographischen Selbstund

Fremdthematisierung
[Theoretical perspectives on biographical conceptions about

development]. In
J. Matthes (Ed.), Lebenswelt und soziale Probleme.

Verhandlungen des 20.
Deutschen Soziologentages (pp. 502-520). Frankfurt: Campus.
Krueger, 1., & Heckhausen, 1. (1991). Perception of growth

and decline on five person
ality dimensions across the adult life span. Manuscript

submitted for publication,
Max Planck Institute for Human Development and Education,

Berlin, FRO.
Lewinsohn, P. M., Mischel, W., Chaplin, W., & Barton, R.

(1980). Social competence
and depression: The role of illusory self-perceptions.

Journal of Abnormal
Marini, M. M. (1984). Age and sequencing norms in the the

transition to adulthood.
Social Forces, 63, 229-244.
Mayer, K.-U. (1986). Structural constraints on the life

course. Human Development, 29,
163-170.
Modell, 1., Furstenberg, F. F., Jr., & Hershberg, T.

(1976). Social change and transition
to adulthood in historical perspectives. Journal of Family

History, /, 7-32.
Neugarten, B. L. (1968). The awareness of middle age. In

B. L. Neugarten (Ed.),
Middle age and aging (pp. 93-98). Chicago: University of

Chicago Press.
Neugarten, B. L., Moore, J. W., & Lowe, 1. C. (1965). Age

norms, age constraints, and
adult socialization. American Journal of Sociology, 70,

710-717.
Norman, W. T. (1963). Toward an adequate taxanomy of

personality attitudes: Repli
cated factor structure in peer nomination personality

ratings. Journal of Abnormal
O'Oorman, H. J. (1980). False consciousness of kind:

Pluralistic ignorance among the
aged. Research on Aging, 2, 105-128.
Piaget, 1. (1985). The equilibration of cognitive

structures: The central problem of
intellectual development. Chicago: University of Chicago

Press.
Riley, M. W., & Riley, J. W., Jr. (1986). Longevity and

social structure: The added
years. Daedalus, 115,51-75.

Rothbaum, F., Weisz, 1. R., & Snyder, S. S. (1982).
Changing the world and changing
the self: A two-process model of perceived control.

Journal of Personality and Social
Salthouse, T. A. (1985). A theory of cognitive aging.

Amsterdam: North Holland.
Schulz, R., & Decker, S. (1985). Long-term adjustment to

physical disability: The role
of social support, perceived control, and self-blame.

Social Psychology, 48, 1 J 62-1172.
Schulz, R., Heckhausen, J., & Locher, J. (in press). Adult

development, control, and
adaptive functioning. Journal of Social Issues.
Snyder, C. R., & Frornkin, H. L. (1980). The psychology of

uniqueness. New York:
Plenum.
Snyder, M. L., Stephan, W. G., & Rosenfield, D. (1978).

Attributional egotism. In
1. H. Harvey, W. Ickes, & R. F. Kidd (Eds.), New directions

in attribution research
(Vol. 2, pp. 91-117). Hillsdale, NJ: Erlbaum.
Steinberg, L., & Silverberg, S. B. (1986). The

vicissitudes of autonomy in early adoles
cence. Child Development, 57,841-851.
Taylor, S. E. (1989). Positive illusions: Creative

self-deception and the healthy mind.
Taylor, S. E., & Lobel, M. (1989). Social comparison

activity under threat: Downward
evaluation and upward contacts. Psychological Review, 96,
569-575.
Taylor, S. E., Wood, J. V., & Lichtman, R. R. (1984).

Attributions, beliefs about con
trol, and adjustment to breast cancer. Journal of

Personality and Social Psychology,
46, 489-502.
Wills, T. A. (1981). Downward comparison principles in

social psychology. Psycholog
ical Bulletin, 90, 245-271.
Wood, 1. V. (1989). Theory and research concerning social

comparison of personal
attributes. Psychological Bulletin, 106(2), pp. 231-248.
Zuckerman, M. (1979). Attribution of success and failure

revisited, or: The motiva
tional bias is alive and well in attribution theory.

Journal of Personality, 47, 245-287. PERSONAL CONTROL OVER
DEVELOPMENT: SOME DEVELOPMENTAL IMPLICATIONS OF
SELF-EFFICACY lochen BrandtsHidter The present
contribution explores the various ways in which
selfpercepts of efficacy may contribute to optimal
development and successful aging. Evidence from a larger
cross-sequential investigation is presented supporting the
view that self-referential beliefs of efficacy and
control are of key importance in maintaining an optimistic
perspective on personal development during middle and later
adulthood. Notions of control and efficacy, however,
cannot fully account for the mechanisms of coping with
uncontrollable events and irreversible losses that
typically cumulate in later phases of life. It is argued
that besides self-percepts of efficacy, the capability or
readiness to disengage from blocked developmental options
and to flexibly readjust one's developmental goals is an
important factor that reduces the risk of dissatisfaction
and depression in later life. Empirical findings indicate
that this second, accommodative mode of coping becomes
increasingly dominant in later adulthood.
The present contribution attempts to highlight some

implications that notions of
personal control and self-efficacy have for issues of
optimal development and
successful aging. Such implications become readily apparent

from an action per
spective of development that focuses on the individuals'

active contribution in
shaping their personal development and circumstances of

living (cf. Bandura,
1981; Brandtstlidter, 1984). From this perspective,

development over the life
span has to be conceived as a process that to a large

extent depends on the devel
oping subject's self-referential cognitions, evaluations

and actions. Life is a
history of gains and losses in diverse areas of life and

functioning (cf. Heck
hausen, Dixon, & Baltes, 1989), but also a history of more

or less successful at
tempts to keep this balance of gains and losses favorable.

We are of course not
the omnipotent producers of our development, as Bandura

(1982a) has reminded
us in his notion of "chance encounters." Schopenhauer, in

his counsels and
maxims concerning the wisdom of life, expressed this

insight as follows: " ... the
course of a man's life is in no wise entirely of his own

making; it is the product
of two factors-the series of things that happened, and his

own resolves in regard
to them, and these two are constantly interacting upon and

modifying each other"
(Schopenhauer, 1851/1951, p. 84). Extending notions of

control and self-efficacy to the developmental domain
spawns many interesting questions: e.g., how do

self-beliefs of efficacy and con
trol over development influence the person's construction

of developmental gains
and losses over the life span? How and to what extent do
such self-referential
beliefs determine the person's motivation of readiness to

counteract anticipated
developmental losses? What is their functional role in

coping with critical life
transitions, especially with those partly irreversible and

uncontrollable losses that
characterize the later phases of life? The following

sections will address these
questions. PERCEIVED EFFICACY IN INSTRUMENTAL COPING

EFFORTS: AN ACTION· THEORETICAL PERSPECTIVE The
action-theoretical model shown in Figure 1 may help to
elucidate the
functional role of self-efficacy beliefs in the context of

personal control over
development (see also Brandtstadter, 1989; Brandtstadter &

Renner, 1991). The
scheme roughly delineates cognitive, affective and

behavioral factors involved in
the process of coping with developmental losses and

deficits. More specifically,
it portrays a first phase or mode of coping where the

individual actively tries to
alter his or her behavior or developmental circumstances so

that they match with,
or are assimilated to, his or her developmental goals, life

themes and desired
"possible selves" (Markus & Nurius, 1986). We have denoted

this stage in the
coping process as the assimilative mode, in contrast to a

second, accommodative
circumstances (cf. Brandtstadter, 1989; Brandtstadter &

Renner, 1990). This
second mode or phase of coping, and its relation to

self-referential efficacy
beliefs, will be considered more closely in a later

section. As shown in Figure I, active-assimilative modes
of coping are initially moti
vated by a perceived discrepancy or mismatch between

factual and desired devel
opmental prospects (the corresponding self-monitoring

process is addressed in
the Components I and 2 of the model). The ensuing

affective and behavioral
consequences now will critically depend on the individual's

appraisal of his or
her potentials for altering this unsatisfactory state. The

different branching points
in the scheme (4,7,9) denote different courses of events

(for clarity, the branching
points are conceived as alternatives: yes/no, +/-). If

personal control potentials
for altering the situation are seen as sufficient (5,+),

the person will take
measures to ameliorate his or her developmental

circumstances and prospects (6).
If, in contrast, personal potentials for action are seen as

deficient (5,-), this will
temporarily block development-related change efforts, but

may not immediately
bring about a state of helplessness or depression. Rather,

it seems plausible to
assume that at this stage of the coping episode,
individuals will first try to
augment their capacities and resources of action by

searching for relevant
Figure 1 The assimilative mode of coping with developmental

losses and deficits. Different active-assimilative
strategies of coping are denoted by the paths < 1 ~ 2 (-)
~ 4 ~ 5 (+) ~ 6 >, < I ~ 2 (-) ~ 4 ~ 5 (-) ~ 7 (+) ~ 8 ~ 4
~ 5 (+) ~ 6 >, < 1 ~ 2 (-) ~ 4 ~ 5 (-) ~ 7 (-) ~ 9 ~ 10 >.
Efficacy beliefs intervene at the branching points of the
scheme, which are linked with characteristic emotional
states (according to this model, resignation and
depression mark the shift from assimilative to
accommodative processes of coping; see text for further
explanations). v Evaluation of personal developmenta l
prospects and conditions 2 Is developmental situation
satisfactory? .3 Stati c orientation, low corrective
tendency CONTENTMENT, SATISFACTION DISAPPOINTMENT,
ANNOYANCE Appraisal of potentials and limitations for
corrective action 4 CONFIDENCE, TENSION 6
Development-related change elforts (planning, executive
control) 5. Is personal control potential sufficient? B:
Enhancement ol personal control potential SELF-DOUBTS,
WORRY 7 Is Improvement of personal control potential
possible ? DEPRESSION, RESIGNATION GRATITUDE, HOPE Are
external supports and resources available? 9J Iff
Involvemen t of external suppor t systems (proxy control)
Inhibition of Instrumental efforts, activation of
accommodative tendencies 11
information, acquiring new skills, and so on-provided that

such options are seen
as available (5,-; 7,+). If such additional efforts turn

out as unavailing, the indi
vidual may finally try to engage external support (7,-; 9,

10) or use some kind of
proxy control (Bandura, I 982b) to manage the problem.

Feelings of helplessness and hopelessness will arise when
the different
options for meliorative intervention depicted at various

levels of the scheme seem
exhausted (2,-; 5,-; 7,-; 9,-). Obviously, self-percepts of

efficacy and control,
which of course themselves will be influenced by

experiences of success or fail
ure in the assimilative phase of coping, function as

differential moderating
parameters at the different branching points of the

modeled sequence. Persons
harboring doubts about their potentials and resources to

prevent personally un
desired developmental outcomes and anticipated

developmental losses will gen
erally be less prone to engage in active efforts to

ameliorate their developmental
prospects or, if confronted with obstacles, to enhance

their control potentials. In
brief, the model shows how self-percepts of low efficacy

pave the way into de
pression, or, conversely, how self-efficacy beliefs may

enhance maintaining or
regaining a positive, optimistic outlook on personal

development. PERCEIVED CONTROL OVER DEVELOPMENT AND
QUALITY OF LIFE PERSPECTIVES Against the backdrop of the
action-theoretical model depicted above, it
seems plausible to assume that differences in perceived

self-efficacy or control
over personal development should be closely related to

indicators of optimal
development and successful aging, and that strong percepts

of efficacy will, at
least in the long run, payoff in a more favorable balance

of developmental gains
and losses. In the following, I will present some selected

findings that lend
support to this general assumption. Before turning to

results, let me briefly
describe the research project where these findings come

from. The project I focuses on issues of personal
self-regulation of development in
adulthood. Our panel involves over 1,200 participants in

the age range from 30 to
60 years. Within a cross-sequential research design which

combines cross
sectional and longitudinal comparisons, structured

questionnaires were used to
assess various facets of the individual's appraisal of his

or her own develop
mental situation. The ratings were done with respect to

different goal dimensions
(e.g., subjects were asked to rate the perceived importance

of goal, the subjective
distance from goal, the extent to which attainment of goals

depends on personal
efforts). We also asked our participants to describe their

feelings with regard to
their past and future development on selected adjective

scales. From the basic rat
ings, various global indicators of development-related

perceptions, beliefs and
I "Entwicklungserleben und Entwicklungskontrolle in

Partnerschaften," funded by the
German Research Foundation. The author is grateful to

Bernhard Baltes-GlHz, Werner
Greve, GUnter Krampen, Gerolf Renner and Dirk Wippermann,

who assisted in various
phases of this research.
action tendencies were derived. Further measures were

included to assess inter
individual differences in personality traits, generalized
control beliefs, marital
adjustment, and life satisfaction. The longitudinal

replications were separated by
a two-year interval; to date, three waves have been

completed which together
span a longitudinal interval of four years (for more

detailed descriptions of the
research approach, see Brandtstadter & Baltes-Gotz, 1991).

Development-Related Control Beliefs and Perceived
Developmental Deficits A first series of analyses centered
on the question of how self-percepts of
control over development relate to perceived developmental

deficits, or distances
from developmental goals. Figure 2 summarizes the findings

for the first wave,
which were found to be stable across all longitudinal

replications. The figure re
lates the distance ratings for 17 different goal dimensions

of personal develop
ment ("How far are you presently from achieving this

goal?") to an index var
iable of personal control over development (peD; the two

profiles shown in
Figure 2 compare subgroups equal to or above and below the

median of the PCD
variable). The PCD index was aggregated from ratings

concerning the perceived
impact of controllable or autonomous and uncontrollable or

heteronomous fac
tors on personal development; validation studies have

confInned its usefulness as
a measure of perceived control or development-related

self-efficacy (see Brandt
stadter, 1989). Across all goal dimensions, individuals
scoring low in perceived
control over development report significantly higher

developmental deficits.
There is also a control by goal dimensions interaction:

Subjects having self
percepts of low control over development were found to

report greater deficits
above all on dimensions related to health, occupational

efficiency, prosperity, and
intellectual efficiency. A quite similar pattern of

findings emerges when we take
as a dependent variable the self-attributed potential for

further developmental
progress on the different dimensions (see Brandtstadter,

Krampen, & Greve,
1987). These results indicate that subjects having

self-percepts of high personal
control over development give a more favorable account of

their actual develop
mental situation. At the same time, they also see a greater

latitude for improving
their situation through detennined efforts. Since no

interactions involving age or
gender were observed, it seems that we can generalize our

findings across these
factors. Emotional hnplications of Perceived Control Over

Development As may already be extrapolated from the
previous findings, self-percepts of
control over development should go with an optimistic and

zestful outlook on
personal development. Persons with self-confident

action-outcome expectancies
should be less threatened by aversive developmental

prospects, and less vulner
able to feelings of despondency and depression.

Parenthetically, we may note that
such feelings may not only result from distinct "bad

events" (Peterson &
Seligman, 1984) that are perceived as irreversible, but

also from the experience
of a gradual worsening in the balance of gains and losses

with advancing age (cf.
Heckhausen, Dixon, & Baltes, 1989). To some extent, the

emotional implications
of perceived control potentials or deficits are already

revealed by analyzing the
conceptual structure of emotion tenns (e.g.,

Brandtstadter, 1985). 2 3 4 (1) health (2) emotional
stability (3) wisdom, mature understanding of life (4)
self-esteem (5) social recognition (6) occupational
efficiency (7) assertivness, self assurance (8)
harmonious partnership (9) empathy (10) personal
independence (11) family security (12) prosperity,
comfortable standard of living (13) intellectual
effiCIency (14) self-development, actualization of
personal potential (IS) physical fitness (16) satisfying
friendship (17) commitment to ideals 5
Figure 2 Perceived distance from developmental goals as a

function of personal control over development (PCD). The
profiles of means compare subgroups equal to or above (e)
and below (0) the median of the PCD index (Brandtstadter,
Krampen, & Greve, 1986). In our research, the postulated
cognitive-emotional relationhips were borne
out in many convergent observations. As a paradigm case, we

may consider the
mUltiple regression of perceived control over development

(as measured by the
PCD index, see above) on ratings of emotional attitudes

towards past and future
development. The findings are based on data from the third

wave (1987), but are
highly consistent across all measurement points
(Brandtstadter & Baltes-Gotz,
1991). Inspection of the regression structures shows that

subjects scoring high in
perceived control over development (PCD) tend to describe

themselves proud,
happy, glad, satisfied, grateful when looking back on their

past development;
looking toward their future, they feel more hopefUl, calm,

confident, venturesome.
To get a more detailed picture, we have decomposed here the

aggregated index of
personal control (PCD) into its separate components of

Autonomous Control
and Heteronomous Control over Development (in the

aggregation procedure,
Table I
Multiple Regression of Personal Control Over Development

(PCD). Autonomous
Control Over Development (CDA). and Heteronomous Control

Over Development
(CDH) on Ratings of Emotional Attitude Toward Past and

Future Personal
Development (Multiple Correlations and Regression Structure

Coefficients) Predictor variables PCD CDA CDH
Retrospective Emotions Distressed -.63 -AI 043 Proud .36
044 .02 Happy 045 .39 -.20 Exhausted -.51 -.28 046
Depressed -.62 -.30 .60 Glad .57 .50 -.25 Resigned -.60
-.32 .55 Indifferent -.21 -.30 .4 I Satisfied .50 AS
-.20 Powerless -.61 -.31 .56 Angry -048 -.23 047
Grateful 043 .50 .06 Sad -.62 -.36 049 Prospective
Emotions Hopeful .71 .74 -.15 Calm .55 .52 -.20 Uneasy
-.52 -.30 044 Cheerful .16 .08 -.15 Discouraged -.71
-AI .61 Worried -.16 -.21 -.04 Anxious -.63 -.31 .66
At a loss -.71 -.36 .67 Fearful -.70 -.31 .75
Confident .61 .50 -.34 Troubled -.57 -.30 .56
Venturesome .71 .83 .00 Depressed -.69 -.36 .62 R
.40** AI ** .31** N 757 813 794
Note. ** P.$ .01 (listwise deletion of missing data).
PCD was constructed as the difference between the indexes

of Autonomous and
Heteronomous Control). For Autonomous Control, we find

largely the same pat
tern of relationships as for the global PCD index, whereas

Heteronomous Control
predicts a converse pattern of depression, worry and

despondency (distressed,
exhausted, depressed, discouraged,feaiful, etc.). The

finding of diametrically op
posed correlational patterns for Autonomous and

Heteronomous Control is not
trivial, because these indicators do not form the opposite

poles of a one
dimensional construct, but, on the contrary, were found to

constitute statistically
independent aspects of perceived control. This means, for

example, that people
who consider their personal development as strongly

influenced by external and
uncontrollable factors may nevertheless be convinced that

they can play an active
role in optimizing their balance of gains and losses.

According to our observa
tions, especially older people tend to adopt such a

two-sided view, which
obviously does not easily fit with traditional bipolar

(e.g., internal versus
external) conceptions of control. A very similar pattern

of relationships emerges when we correlate longitudi
nal change scores in the control and emotion variables.

Over a longitudinal inter
val of four years, positive changes in self-percepts of

Autonomous Control over
Development are generally accompanied by shifts toward a

positive evaluation of
personal developmental achievements and prospects, as well

as by other favor
able changes in developmental circumstances and

development-related beliefs
(e.g., increases in general life satisfaction and perceived

marital support, de
creases in various indicators of depression; see

Brandtstadter & Baltes-Gotz,
1991). Conversely, increases in Heteronomous Control over

Development go
with a general worsening of developmental prospects.

Inspection of cross-lagged
panel correlations, however, did not suggest a specific

causal ordering. As al
ready mentioned above, self-percepts of low control over

development should
perhaps be considered as constitutive, rather than as

causal, conditions of
despondency and depression. Self-Percepts of Control and

Subjective Impact of Stressful Events Critical or
stressful life events-especially events or life changes
which,
from the subject's point of view, involve a worsening of

personal developmental
prospects-are generally seen as risk factors for depression

and for a broad
spectrum of pathophysiological changes, even if the

mechanisms that mediate the
documented statistical relationships are yet under dispute

(e.g., Schroeder &
Costa, 1984). Control-theoretical approaches have

underscored the potential buf
fering effect of action-outcome expectancies on this

relationship. As argued
above, such expectancies should determine how much effort

people will expend
and how long they will persist in active-assimilative

coping efforts; furthermore,
self-percepts of efficacy and control should enhance

threat-reducing interpreta
tions of the impending situation (cf. Bandura, 1981;

Lazarus & Folkman, 1984;
Peterson & Seligman, 1984; Rodin, \986; Scheier & Carver,

1985). Generally,
the belief that one cannot improve one's developmental

prospects or cannot attain
personally valued goals is-almost by definition-an

essential feature of help
lessness, depression and alienation. Besides the

assessments already described, we have asked the
participants in
our panel study to report critical or stressful events that

they had experienced in
the recent past and to rate the degree of strain

experienced during the episode. In
a first exploratory step, we correlated these ratings with

age, depressive outlook
on personal development and different measures of

perceived control. Together
with the indicators of Autonomous and Heteronomous Control

over Develop
ment, a German adaptation of Levenson's scales

(IPC-questionnaire; Krampen,
1981) was included in this analysis as a measure of
generalized control beliefs.
Depressive tendencies were measured by an index variable

derived from self
ratings on selected adjective scales (cf. Brandtstadter &

Baltes-Gotz, 1991).
Table 2 shows the relationships for different categories

of events (the data come
from the third wave, 1987; similar findings were obtained

for the earlier waves). First of all, the findings
presented in Table 2 confirm the expected relation
ship of experienced strain in critical episodes with

depressive tendencies; the
event categories of unemployment, conflicts in familial or

occupational contexts,
change in financial state, change in residence and

personal illness seem to stand
out in this relationship. To further trace the effects of

critical events reported for a
given two-year longitudinal interval, we also looked at

changes in depression rat
ings over that interval. Persons afflicted by personal

illness, occupational con
flicts and by conflicts with family members or friends

showed a significantly
greater increase in depressive tendency (and a

corresponding decrease in life
satisfaction) than the complementary subsample of

individuals not reporting the
given event. For the other categories of events,

longitudinal changes mostly
pointed in a similar direction but-partly due to

restrictions in sample size for the
given event type-fell short of significance (noticeable
exceptions are the cate
gories pregnancy/birth of child, change of residence and

occupational change,
which tended to go with a reduction in depressive

tendencies over the
corresponding longitudinal interval). The focus of

interest here, of course, is on the associations of
experienced
strain with indicators of control and self-efficacy. As a

general tendency, control
beliefs indicating low self-efficacy or an extemallocus of

control (Heteronomous
Control, IPC-Powerful others, IPC-Chance control) seem

positively related to the
strain experienced in a given critical episode. This

relationship appears most
clear-cut for personal illness, unemployment, and

occupational conflicts. Inter
estingly, these data also hint that in specific cases,

self-percepts of control over
development may go with greater emotional strain in

critical situations. A plaus
ible post factum explanation would be that persons

attributing themselves a high
degree of control over their development may to a greater

extent feel personally
responsible for failures and setbacks. Within a

hierarchical regression format, we further looked for
moderating
effects of control beliefs on the relationship between

exposure to critical events
and emotional strain, using simple and autoregressively

residualized differences T a b l e 2 S u b j e c t i v e I
m p a c t o f S t r e s s f u l E v e n t s : R e l a t i o
n s W i t h A g e , D e p r e s s i v e O u t l o o k , a n
d M e a s u r e s o f C o n t r o l T y p e o f E v e n t a
R e f e r e n c e V a r i a b l e 1 2 3 4 5 6 7 8 9 1 0 1 1
1 2 ( 8 2 ) ( 2 5 4 ) ( 4 2 6 ) ( 2 0 4 ) ( 2 9 ) ( 4 0 ) (
1 9 5 ) ( 1 0 9 ) ( 1 9 1 ) ( 3 8 8 ) ( 1 9 3 ) A g e . 0 1
. 0 3 . 1 8 * * . 0 0 . 2 4 . 2 2 . 0 3 . 0 2 . 0 1 . 0 1 .
0 1 . 0 4 D e p r e s s i v e o u t l o o k ( D E P ) . 3 1
* * . 2 7 * * . 1 5 * * . 1 7 * * . 0 8 . 4 4 * * . 3 6 * *
. 1 0 . 3 3 * * . 3 1 * * . 3 8 * * . 1 7 * * A u t o n o m
o u s c o n t r o l o v e r d e v e l o p m e n t ( C D A )
. 0 7 . 0 0 . 1 2 * . 1 1 . 1 2 . 2 2 . 2 2 * * . 1 3 . 0 2
. 1 6 * * . 0 9 . 0 2 H e t e r o n o m o u s c o n t r o l
o v e r d e v e l o p m e n t ( C D H ) . 0 2 . 2 4 * * . 2
2 * * . 0 7 . 0 4 . 3 5 * . 1 4 . 2 8 * * . 2 7 * * . 0 9 .
0 2 . 0 8 * P e r s o n a l c o n t r o l o v e r d e v e l
o p m e n t ( P C D ) . 0 0 . 1 6 * . 0 6 . 0 3 . 1 1 . 0 9
. 0 3 . 0 7 . 1 7 * . 0 5 . 0 3 . 0 5 I n t e r n a l i t y
( I P C I ) . 2 6 * . 0 7 . 0 2 . 1 2 . 3 0 . 0 4 . 0 2 . 0
9 . 1 4 . 0 4 . 0 1 . 0 6 P o w e r f u l o t h e r s ( I P
C P ) . 1 0 . 2 4 * * . 0 9 . 0 8 . 0 6 . 3 8 * . 1 8 * . 2
3 * . 2 3 * * . 0 1 . 1 6 * . 1 2 * * C h a n c e c o n t r
o l ( I P C C ) . 1 0 . 2 2 * * . 1 8 * * . 0 3 . 0 9 . 2 2
. 2 8 * * . 0 8 . 2 4 * * . 0 8 . 1 9 * * . 1 0 * * N o t e
. a 1 : C h a n g e i n r e s i d e n c e , 2 : P e r s o n
a l i l l n e s s , 3 : I l l n e s s o f f a m i l y m e m
b e r , 4 : D e a t h o f f a m i l y m e m b e r , S : P r
e g n a n c y / b i r t h o f c h i l d , 6 : U n e m p l o
y m e n t , 7 : C h a n g e i n f i n a n c i a l s t a t e
, 8 : O c c u p a t i o n a l c h a n g e , 9 : O c c u p a
t i o n a l c o n f l i c t s , 1 0 : C o n f l i c t s w i
t h f a m i l y m e m b e r s o r f r i e n d s , 1 1 : O t
h e r s t r e s s f u l e v e n t ( o p e n c a t e g o r y
) , 1 2 : S u m o f r e p o r t e d c r i t i c a l e v e n
t s . N u m b e r s i n p a r e n t h e s e s r e f e r t o
f r e q u e n c y o f t h e g i v e n e v e n t w i t h i n
t h e s a m p l e ( N = 9 9 8 ) . • • p < . 0 1 ; . p < . 0
5 . i l
between preand post-event depression ratings (obtained at

the second and third
occasion of measurement, respectively) as dependent

variables. Space considera
tions allow for only a condensed overview here. Observed

moderation effects
were highly specific for the type of event considered. In

part, the effects were
consistent with the presumed buffering effect of perceived
control; for example,
it was found that higher scores on IPC-internality and on

Autonomous Control
over Development dampen the negative emotional impact of

occupational con
flict. Such effects came out more clearly when pertinent

domain-specific ratings
of perceived control over development were considered

(considering, e.g., the
emotional impact of personal illness, health-related

control beliefs seem to have a
stronger moderating effect than general indicators of

perceived control). The
global picture of our findings, however, does not support

the unconditional
conclusion that self-percepts of control and self-efficacy

are under all circum
stances contributive to emotional resilience and effective

coping. For example,
our results hint that the emotional impact of the event

"death of family member"
is not mitigated, but rather aggravated by perceived

control or self-efficacy. This
indicates that active-assimilative efforts at control may

be dysfunctional in situa
tions of loss which are factually irreversible. It is

possible, then, that the arguments considered so far, which
largely cen
tered on salutory effects of self-efficacy, give only an

incomplete grasp of the
factors that help individuals to cope with life crises and

make them less vulner
able to depression in situations of loss. In the

following, I will briefly sketch the
outlines of a more comprehensive theoretical perspective.

BROADENING THE THEORETICAL SCOPE: TENACIOUS PURSUIT AND
FLEXIBLE ADJUSTMENT OF DEVELOPMENTAL GOALS To sum up,
there are good empirical and theoretical reasons supporting
the
view that self-referential beliefs of efficacy and control

are of key importance in
maintaining a positive and optimistic outlook on personal

development in middle
and later adulthood. But-as intimated above-I think that

this is only one ele
ment of a more complex story. We have a thorough

theoretical understanding of
the different ways in which self-percepts of efficacy may

contribute to effective
coping, and why persons entertaining doubts in their

control potentials are more
vulnerable to helplessness and depression when faced with

obstacles or failures.
From a control-theoretical or learned helplessness

perspectives, however, it is by
far less clear how people manage to recover from

resignation and depression, es
pecially in cases where they are confronted with permanent

and definitely
irreversible loss. Accommodative Modes of Coping To

approach this question theoretically, we should recognize
that a discrep
ancy between actual and desired developmental prospects may

be handled in two
basically different ways: on the one hand, the situation

may be transformed-in
the sense of assimilative efforts given above-to correspond

more closely to per
sonal goals and aspirations; on the other hand, the
discrepancy may be neutral
ized by alterations in the system of personal goals,

aspirations and evaluative
cognitions that make the previously aversive situation more

acceptable. It follows that besides self-referential
beliefs of control and efficacy, the cap
ability or readiness to disengage from thwarted

developmental options and to
flexibly revise and readjust one's developmental goals and

life design may be an
important (and hitherto largely neglected) factor that

serves to diminish the
impact of aversive and stressful experiences and to reduce

the individual's vul
nerability to depression. Apparently, we are dealing here

with a mode of coping
that is basically different from active-assimilative

efforts. This second mode of
coping, which I have termed as accommodative, should become

predominant to
the extent that active, instrumental efforts to master the

situation seem futile, or
when generalized or specific beliefs of self-efficacy have

been eroded through
repeated unsuccessful attempts. In a certain sense, these

considerations lead us
beyond notions of control and learned helplessness. Note

that accommodative
readjustments of goals, aspirations and evaluative

standards involve to a con
siderable extent reactive, automatic or effortless

processes that do not involve
intentional action. There are of course techniques of
self-management and self
instruction that may be deliberately applied by the

individual in an attempt to
alter aversive cognitions and emotional states (e.g.,

Karoly & Kanfer, 1982; cf.
also the notions of "emotion-focused coping" by Lazarus,

1977, or of "secondary
control" by Rothbaum, Weisz, & Snyder, 1982). The fact,

however, that we can
to some extent modify our cognitive and emotional processes

does not imply that
these processes should be considered as intentionally

controlled actions Uust as
the fact that we can deliberately bring about some bodily

reflexes does not mean
that these reflexes are intentional acts). We cannot give

up our beliefs and com
mitments merely because it seems advantageous to us; if we

could, problems of
depression and despair would presumably not exist. To make

a longer story
short: The paradigm of intentional action does not apply

to the cognitive and
motivational processes that form the basis of intentional

actions (see also Lanz,
1987). Rather, the process of generating palliative

cognitions in aversive situa
tions depends on their availability in a given situation;

research on the mood con
gruence of cognitions has shed some light on the

subpersonal mechanisms
involved (for an overview, see Blaney, 1986). It becomes

apparent at this junc
ture that feelings of hopelessness and helplessness are
not simply the deplorable
end state of unavailing efforts to master a problem, but

may be important param
eters in the shift from assimilative to accommodative

phases of coping. When
faced with factually uncontrollable events and irreversible

losses, persons having
a strong sense of personal control and self-efficacy may

even have greater
difficulties in adjusting their goals and life plans to the

new circumstances (cf.
Brandtstiidter & Renner, t 990; Janoff-Bulman & Brickman,

1982). The accommodative phase of coping involves a
reorganization of goals,
beliefs, aspirations and evaluational standards on

different levels. We assume
that, on a first level, reappraisals of the situation are

activated; as the emotional
evaluation of a given situation largely depends on the

expectations and meanings
(semantic and instrumental) associated with that situation,

changes in subjective
probabilities of specific consequences may have a

palliative effect and enhance
disengagment from blocked goals and action tendencies (cf.

Klinger, 1975). Such
palliative processes critically depend on the subjective

availability of alternative
interpretations, which may be enhanced by exploratory and

ruminative processes
(search for new information, consideration of new

arguments, changing the ana
lytic focus, etc.). Within the boundaries of rationality,
there is usually some lati
tude for alternative interpretations of a given situation

(cf. also Taylor & Brown,
1988); when alternative interpretations are available,

individuals usually tend to
endorse those that are consistent with, and have positive

implications for, their
self-conception and personal view of the world (cf.

Greenwald, 1980). If pallia
tive interpretations are not accessible, individuals may

alter their evaluative
standards and reference points. For example, developmental

losses in old age
may seem more acceptable to individuals who compare

themselves with same
aged rather than with younger persons. Generally, a

situation may appear less
aversive when we contrast it with some worse alternative

or, conversely, avoid
contrasting it with a counterfactual better world (cf.

Kahneman & Miller, 1986;
Taylor, Wood, & Lichtman, 1983; Wills, 1981). Bandura

(I982b, 1989) has
discussed the mediating role of such processes in

selective activation of self
corrective tendencies; here, the emphasis is on their

functional role in the process
of disengaging from barren goals and in revising

developmental perspectives. On
a final level, accommodative processes may eventually

involve a radical shift in
the person's conception of self and the world, comparable

to a paradigm shift in
science. Such more radical changes may be expected in cases
when palliative
reappraisals and readjustments of aspirations are not

available for the individual
(for further theoretical elaborations, see Brandtstiidter

& Renner, 1991). Age-Related Changes in Coping Style: From
Active-Assimilative to Accommodative Modes of Coping The
theoretical perspective outlined provides a conceptual
scheme for deal
ing with some notorious puzzles of successful aging. How

can we explain that
older people do not report a general or dramatic loss in

life satisfaction, even
though they face morbidity, death, and an increasingly

unfavorable balance of
developmental gains and losses (Stock, Okun, Haring, & Wi

Iter, 1983)? Why is
there no increase in the prevalence rates of depression in

old age (cf. Bolla
Wilson & Bleeker, 1989; Kasl & Berkman, 1981; Lewinsohn,

Hoberman, Teri, &
Hauzinger, 1985; Newmann, 1989)? As Blazer (1989, p. 198)

summarizes the
evidence, we should "seriously consider the possibility

that older adults ... may
be ... protected from the development of major or clinical

depression." To illus
trate this point, Figure 3 shows cross-sequential gradients

for self-reported life
satisfaction that were obtained from the first and third

wave of our panel study
(N = 998). The gradients span a cross-sectional range from

30 to 63 years and a
longitudinal interval of four years (as a measure of life

satisfaction, we used an
item selection from scales developed by Neugarten,

Havighurst, & Tobin, 1961).
Apart from a slight decrease in subjective quality of life

for the early phase of
middle adulthood (30-41 years), the general picture from

both cross-sectional and
longitudinal comparisons is one of stability over the age

range considered. Even
for the cohorts approaching later adulthood, there is no

indication of a decrease in
life satisfaction. 16 14 I 3{}-35 D 36-41 ill 42-47

Age Cohorts N 48·53 V 54·59
Figure 3 Subjective quality of life in adulthood:

cross-sectional comparisons and fouryear longitudinal
changes in life satisfaction (occasions 1983 and 1987; age
ranges for cohorts refer to first occasion). From a
control-theoretical or learned helplessness point of view,
these facts
become even more puzzling when we assume that older people

see their life and
personal development as increasingly dependent on factors

beyond personal con
trol, as several studies have documented (cf.

Brandtstlidter & Baltes-Gotz, 1991;
Lachman, 1986; Rodin, 1987). These apparent inconsistencies

may be resolved
when we assume a gradual age-related shift from

assimilative to accommodative
modes of coping. We have investigated this assumption by

means of a question
accommodative modes of coping on a dispositional level.

The instrument com
prises two nearly orthogonal scales that we have denoted as

"Tenacious Goal Pur
suit" and "Flexible Goal Adjustment." Tenacious individuals
cling to goals and
commitments even in the face of obstacles or under high

risk of failure (e.g.,
"When faced with obstacles, I usually double my efforts";

"Even when a situa
tion seems hopeless, J still try to master it"). Flexible

individuals disengage easily 15 1983 1987 L i f e S a t i
s f a c t i o n
Personal Control Over Development
from barren commitments, and try to see the best in

difficult situations (e.g., "I
adapt quite easily to changes in plans or circumstances";

"If I don't get something
I want, I take it with patience"). In spite of their

statistical independence, both
scales correlate consistently and positively with

indicators of successful devel
opment such as optimism, life satisfaction, absence of

depressive tendencies, and
greater resilience in stressful life situations (see

Brandtstadter & Renner, 1990). Figure 4 shows
cross-sectional age gradients for Flexibility and Tenacity
for
a large sample (N = 1,433). These results are obtained by

pooling observations
from our panel study (third wave, 1987; cf. also

Brandtstadter & Renner, 1990)
and several independent investigations in which the

Flexibility and Tenacity
scales were used. Over the age range considered, the age
cohort by coping style
(Flexibility, Tenacity) interaction is highly significant

and clearly corresponds to
the predicted shift from assimilative to accomm'1dative
modes of coping; the
linear correlations of Flexibility and Tenacity with age

are -.23 and +.23,
respectively (the difference between these coefficients is

highly significant).
Inspection of the data for the oldest cohort indicates that

these trends continue
beyond the age of 65. The pattern of clearly opposite

regressions of Flexibility
and Tenacity on age is all the more noteworthy considering

the independence of
the scales (for the total sample, the correlation is .12).

These findings seriously
call into doubt widespread assumptions according to which "

... older people ... cope
in much the same way as younger people" (McCrae, 1982, p.

459); at the same
time, they underscore the importance of dispositional

factors that have been
hitherto largely neglected in coping research (see also

Carver, Scheier, &
Weintraub, 1989). 60 55 c: os .., ~ 1:: 50 0 ..c:

0 U 45 flexible Goal Adjustment Tenacious Goal Pursuil
I :<>30 (192) IT III IV 31-35 36-40 41-45 (116) (167)
(173) v VI vn vm IX 46-50 51-55 56-60 61-65 ~ 66 (213)
(180) (199) (135) (58) Age Cohorts
Figure 4 Age gradients for Tenacious Goal Pursuit and

Flexible Goal Adjustment. 1z (1) c Oz 0 ·u (2) .l!!
In ~ In .l!! (3) :.:::i -1z (1 ) L8 = .57 .18 GD
(F'GA = 2z (2) L8 = .29 .27 GD (FGA = lz) (4) (3) L8
= .02 .37 GD (FGA = Oz) (4) L8 =-.25 .47 GD (FGA =-1
z) (5) L5 =-.53 .57 GD (FGA =-2z) (5) -2z -2z -1z
Oz 1z 2z Sum of goal distances
Figure 5 Conditional regressions of Life Satisfaction (LS)

on subjective developmental deficits (sum index of
perceived distances from 17 developmental goals, GO) for
different levels. Analyses of moderation effects further
support this line of argument. Con
sidering the presumed palliative functions of

accommodative processes, we
should expect that for highly flexible persons, perceived

developmental losses or
deficits should be less detrimental to subjective quality

of life. Figure 5 shows the
conditional regressions of life satisfaction on the sum of

perceived goal distances
over 17 developmental goals for different levels of

Flexibility (N = 885; these
data come from the third wave of our panel study where the
Flexibility scale was
first introduced; see also Brandtstadter & Renner, 1990).

While being signifi
cantly negative (-.43) for the total sample, the

correlation between life satisfac
tion and the sum of perceived distances from developmental

goals is less pro
nounced for higher levels of Flexibility (this moderation

effect is also obtained
when age is statistically controlled). Apparently, the

observed age-related shift
from assimilative to accommodative modes of coping should

not be interpreted
as reflecting an insidious trend toward resignation and

apathy, but rather as a
process that is functional in maintaining an optimistic

perspective, perhaps also a
sense of power and self-efficacy, in old age. At this

juncture, the question may be raised whether the notions
of power and
self-efficacy should not be considered as diametrically
opposed to the accommo
On the one hand, it is certainly true that as long as there

are no reasons to doubt
one's efficacy in attaining personally valued goals, there

is no point for accom
modative adjustments. But if getting what one wants is

central to the concept of
power, it follows that a way to retain a sense of efficacy

may be to adjust one's
Bandura, A. (1981). Self-referent thought: A developmental

analysis of self-efficacy. In
J. H. Flavell & L. Ross (Eds.), Social cognitive

development. Frontiers and possible
futures (pp. 200-239). Cambridge: Cambridge University

Press.
Bandura, A. (1982a). The psychology of chance encounters

and life paths. American
Bandura, A. (1982b). The self and mechanisms of agency. In

J. Suls (Ed.), Psychologi
cal perspectives on the self (Vol. 1, pp. 3-40).

Bandura, A. (1989). Self-regulation of motivation and

action through internal standards
and goal systems. In L. A. Pervin (Ed.), Goal concepts in

personality and social
psychology (pp. 19-85). Hillsdale, NJ: Erlbaum.
Blaney, P. H. (1986). Affect and memory: A review.

Psychological Bulletin, 99, 229
246.
Blazer, D. (1989). Depression in late life: An update.

Annual Review of Gerontology &
Geriatrics, 9,197-215.
Bolla-Wilson, K., & Bleecker, M. L. (1989). Absence of

depression in elderly adults.
Journal of Gerontology: Psychological Sciences, 44, P53-55.
Brandtstiidter, 1. (1984). Personal and social control

over development: Some implicia
tions of an action perspective in life-span developmental

psychology. In P. B. Baltes
& O. G. Brim Jr. (Eds.), Life-span development and

behavior (Vol. 6, pp. 1-32). New
York: Academic Press.
Brandtstiidter, J. (1985). Emotion, Kognition, Handlung:

Konzeptuelle Beziehungen
[Emotion, cognition, action: Conceptual relationships). In

L. H. Eckensberger &
E.-D. Lantermann (Eds.), Emotion und Reflexivitiit

[Emotion and reflexivity)
(pp. 252-261). Miinchen: Urban & Schwarzenberg.
Brandtstiidter, J. (1989). Personal self-regulation of

development: Cross-sequential
analyses of development-related control beliefs and

emotions. Developmental
Brandtstiidter, 1., & Baltes-Gtitz, B. (1991). Personal

control over development and
quality of life perspectives in adulthood. In P. B. Baltes

& M. M. Baltes (Eds.), Suc
cessful aging. Perspectives from the behavioral sciences

(pp. 197-224). New York:
Brandtstiidter, 1., Krampen, G., & Greve, W. (1987).

Personal control over develop
ment: Effects on the perception and emotional evaluation

of personal development in
adulthood. International Journal of Behavioral Development,

10,99-120.
Brandtstiidter, 1., & Renner, G. (1990). Tenacious goal

pursuit and flexible goal adjust
ment: Explication and age-related analysis of assimilative

and accommodative strate
gies of coping. Psychology and Aging, 5,58-67.
Brandtstiidter, J., & Renner, G. (1991). Coping with

discrepancies between aspirations
and achievements in adult development: A dual-process

model. In L. Montada,
S.-H. Filipp, & M. R. Lerner (Eds.), Crises and experiences

of loss in adulthood.
Hillsdale, NJ: Erlbaum, in press.
Carver, C. S., Scheier, M. F., & Weintraub, S. K. (1989).

Assessing coping strategies: A
theoretically based approach. Journal of Personality and

Social Psychology, 56,267
283.
Greenwald, A. G. (1980). The totalitarian ego: Fabrication

and revision of personal
history. American Psychologist, 35,603-618.
Heckhausen, 1., Dixon, R. A., & Baltes, P. B. (1989). Gains

and losses in development
throughout adulthood as perceived by different adult age

groups. Developmental
Janoff-Bulman, R., & Brickman, P. (1982). Expectations and

what people learn from
failure. In N. T. Feather (Ed.), Expectations and actions:
Expectancy-value models in
psychology (pp. 207-237). Hillsdale, NJ: Erlbaum.
Kahneman, D., & Miller, D. T. (1986). Norm theory:

Comparing reality to its alterna
tives. Psychological Review, 93, 136-153.
Karoly, P., & Kanfer, F. H. (Eds.). (1982). The psychology

of self-management. From
theory to practice. New York: Plenum.
Kasl, S. V., & Berkman, L. F. (1981). Some psychosocial

influences on the health status
of the elderly: The perspective of social epidemiology. In

J. L. McGaugh &
S. B. Kiesler (Eds.), Aging: Biology and behavior (pp.

345-377). New York:
Academic Press.
Klinger, E. (1975). Consequences of commitment to and

disengagement from
incentives. Psychological Review, 82, 1-25.
Krampen, G. (l981).IPC-Fragebogen zu Kontrolluberzeugungen

rIPC-questionnaire of
Perceived Controll. Gtittingen: Hogrefe.
Lachman, M. E. (1986). Locus of control in aging research.

Journal of Psychology and
Aging, /, 34-40.
Lanz, P. (1987). Menschliches Handeln zwischen Kausalitiit

und Rationalitiit rHuman
action between causality and rationalitYI. Frankfurt a.

M.: Atheniium.
Lazarus, R. S. (1977). Cognitive and coping processes in

emotion. In A. Monat &
R. S. Lazarus (Eds.), Stress and coping (pp. 145-158). New
York: Columbia
University Press.
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal

Springer.
Lewinsohn, P. M., Hoberman, H. M., Teri, L., & Hautzinger,

M. (1985). An integrative
theory of depression. In S. Reiss & R. R. Bootzin (Eds.),

Behavioral issues in
behavior therapy (pp. 331-362). New York: Academic Press.
Markus, H., & Nurius, P. (1986). Possible selves. American

Psychologist, 4/, 954-969.
McCrae, R. R. (1982). Age differences in the use of coping

mechanisms. Journal of
Gerontology, 37, 454-460.
Neugarten, B. L., Havighurst, R. 1., & Tobin, S. S.

(1961). The measurement of life
satisfaction. Journal of Gerontology, 16, 134-143.
Newmann, 1. P. (1989). Aging and depression. Psychology and

Aging, 4, 150-165.
Peterson, C., & Seligman, M. E. P. (1984). Causal

explanations as a risk factor for
depression: Theory and evidence. Psychological Review, 91,

347-374.
Rodin, J. (1986). Health, control and aging. In M. M.

Baltes & P. B. Baltes (Eds.), The
psychology of control and aging (pp. 139-165). Hillsdale,

NJ: Erlbaum.
Rodin, J. (1987). Personal control through the life course.

In R. P. Abeles (Ed.), Life
span perspective and social psychology (pp. 103-119).
Rothbaum, F., Weisz, J. R., & Snyder, S. S. (1982).

Changing the world and changing
the self. A two-process model of perceived control.

Journal of Personality and Social
Scheier, M. F., & Carver, C. S. (1985). Optimism, coping.


Psychology, 4. 219-247.
Schopenhauer. A. (1951). Essays from the parerga and

paralipomena. London: Allen
and Unwin (Original: 1851).
Schroeder. D. H .• & Costa, P. T. Jr. (1984). The

influence of life stress on physical
illness: Substantive effects of methodological flows?

Social Psychology. 46. 853-863.
Stock. W. A .• Okun. M. A .• Haring. M. 1.. & Wilter. R.

A. (1983). Age and subjective
well-being: A meta-analysis. In R. 1. Light (Ed.),

Evaluation studies: Review annual
(Vol. 8. pp. 279-302). Beverly Hills. CA: Sage.
Taylor. S .• & Brown. J. D. (1988). Illusion and

well-being: A social psychological per
spective of mental health. Psychological Bulletin. 103.

193-210.
Taylor. S. E .• Wood. J. U .• & Lichtman. R. R. (1983). It

could be worse: Selective
evaluation as a response to victimization. Journal of

Social Issues, 39, 19-40.
Wills. T. A. (1981). Downward comparison principles in
social psychology.
Psychological Bulletin. 9 J • 245-271. This page

intentionally left blank
III Self-Efficacy, Stress, and Emotions
Arnow, B. A., Taylor, C. B., Agras, W. S., & Teich, M. J.

(\ 985). Enhancing agora
phobia treatment outcome by changing couple communication

patterns. Behavior
Therapy, /6,452-467.
Bandura, A. (\ 969). Principles of behavior modification.

New York: Holt, Rinehart, &
Winston.

Bandura, A. (1978a). On paradigms and recycled ideologies.

Research, 2, 79-103.
Bandura, A. (1978b). Reflections on self-efficacy.

Advances in Behaviour Research and
Therapy, J, 237-269.
Bandura, A. (1982). The assessment and predictive

generality of self-percepts of
efficacy. Journal of Behavior Therapy and Experimental

Psychiatry, /3, 195-199.
Bandura, A. (1984). Recycling misconceptions of perceived

self-efficacy. Cognitive
Bandura, A. (1986). The explanatory and predictive scope

of self-efficacy theory.
Journal of Clinical and Social Psychology, 4, 359-373.

Anxiety Research, J, 77-98.
Bandura, A. (1991). Human agency: The rhetoric and the

reality. American Psycholo
gist,46,157-162.
Bandura, A., & Adams, N. E. (J 977). Analysis of

self-efficacy theory of behavioral
change. Cognitive Therapy and Research, J, 287-308.
Bandura, A., Adams, N. E., & Beyer, J. (1977). Cognitive

processes mediating behavior
change. Journal of Personality and Social Psychology, 35,

125-139.
Bandura, A., Adams, N. E., Hardy, A., & Howells, O. (\

980). Tests of the generality of
self-efficacy theory. Cognitive Therapy and Research,

4,39-66.
Bandura, A., Reese, L., & Adams, N. E. (1982).

Microanalysis of action and fear
arousal as a function of differential levels of perceived

Personality and Social Psychology, 43, 5-2l.
Barlow D. H. (1988). Anxiety and its disorders. New York:

Guilford.
Beck, A. T. (1976). Cognitive therapy and the emotional

disorders. NY: International
Universities Press.
Beck, A. T., Emery, G., & Greenberg, R. L. (1985). Anxiety

disorders and phobias: A
cognitive perspective. NY: Basic Books.
Biran, M., & Wilson, G. T. (198 I). Treatment of phobic

disorders using cognitive and
exposure methods: A self-efficacy analysis. Journal of

Consulting and Clinical

Bourque P., & Ladouceur R. (1980). An investigation of
various performance-based
treatments with agoraphobics. Behaviour Research and

Therapy, 18, 16 I -170.
Craske, M. G., Rapee, R. M., & Barlow, D. H. (1988). The

significance of panic
expectancy for individual patterns of avoidance. Behavior

Therapy, 19, 577-592.
Crowe, M., Marks, I., Agras, W., & Leitenberg, H. (1972).

Time-limited desensitiza
tion, implosion, and shaping for phobic patients: A

crossover study. Behavior
Research and Therapy, 10, 319-328.
Emmelkamp, P. M. G., & Felten, M. (1985). The process of

exposure in vivo: Cognitive
and physiological changes during treatment of acrophobia.

Behaviour Research and
Therapy, 23,219-223.
Gauthier, 1., & Ladouceur, R. (1981). The influence of

self-efficacy reports on perform
ance. Behavior Therapy, 12,436-439.
Gelder, M. (1977). Behavioral treatment of agoraphobia.

Some factors which restrict
change after treatment. In J. C. Boulougouris & A. D.

Rabavilas (Eds.), The treatment
of phobia and obsessil'e compulsive disorders (pp. 7-12).

NY: Pergamon.
Kinney, P. 1., & WiJliams, S. L. (1988). Accuracy of fear

inventories and self-efficacy
scales in estimating agoraphobic behavior. Behaviour

513-518.
Ladouceur, R. (1983). Participant modeling with or without
cognitive treatment of pho
bias. Journal of Consulting and Clinical Psychology,

51,942-944.
Lang, P. 1. (1971). The application of psychophysiological

methods to the study of
psychotherapy and behavior modification. In A. E. Bergin &

S. L. Garfield (Eds.),
Handbook of psychotherapy and behavior change (pp. 75-125).

New York: Wiley.
Leitenberg, H. (1976). Behavioral approaches to the

treatment of neuroses. In H.
Leitenberg (Ed.), Handbook of behavior modification and

behavior therapy (pp. 124167). Englewood Cliffs, NJ:
Prentice-Hall.
Mathews, A. M., Gelder, M. G., & Johnston, D. W. (\981).

Agoraphobia: Nature and
treatment. New York: Guilford.
Meyer, V., & Gelder, M. G. (1963). Behaviour therapy and

phobic disorders. British
Journal of Psychiatry, 109,19-28.
Mowrer, O. H. (1960). Learning theory and behavior. New

York: Wiley.
Rachman, S. (1976). The passing of the two-stage theory of

fear and avoidance: Fresh
possibilities. Behaviour Research and Therapy, 14, 125-131.
Rosenthal, T. L., & Bandura, A. (1978). Psychological

modeling: Theory and practice.
In S. L. Garfield & A. E. Bergin (Eds.), Handbook of

psychotherapy and behavior
change (pp. 621-658). New York: Wiley.
Seligman, M. E. P., & Johnston, J. C. (1973). A cognitive

theory of avoidance learning. In F. J. McGuigan & D. B.
Lumsden (Eds.), Contemporary approaches to condition
ing and learning (pp. 69-110). Washington, DC: Winston &

Sons.
Southworth, S., & Kirsch, I. (1988). The role of

expectancy in exposure-generated fear
reduction in agoraphobia. Behaviour Research and Therapy,

26, 113-120.
TeIch, M. J., Agras, W. S., Taylor, C. B., Roth, W. T., &

Gallen, C. C. (1985). Combin
ed pharmacological and behavioral treatment for

agoraphobia. Behaviour Research
and Therapy, 23,325-335.
TeIch, M. J., Bandura, A., Vinciguerra, P., Agras, A., &

Stout. A. L. (1982). Social
demand for consistency and congruence between self-efficacy

and performance.
Behavior Therapy, /3,694-701.
Williams, S. L. (1985). On the nature and measurement of

agoraphobia. Progress in
Behavior Modification, /9, 109-144.
Williams, S. L. (1987). On anxiety and phobia. Journal of

Anxiety Disorders, /, 161180.
Williams, S. L. (\ 988). Addressing misconceptions about

phobia, anxiety, and self
efficacy: A reply to Marks. Journal of Anxiety Disorders,

2,277-289.
Williams, S. L. (1990). Guided mastery treatment of

agoraphobia: Beyond stimulus
exposure. Progress in Behavior Modification, 26,89-121.
Williams, S. L. (1991 ). Unpublished data.
Williams, S. L., Dooseman, G., & Kleifield, E. (1984).

Comparative effectiveness of
guided mastery and exposure treatments for intractable
phobias. Journal of Consult
ing and Clinical Psychology, 52,505-518.

Generalization of therapeutic changes in agoraphobia: The
role of perceived self-efficacy. Journal of Consulting and
Williams, S. L., & Kleifield, E. (\985). Transfer of

behavioral change across phobias in
multiply phobic clients. Behavior Modification, 9, 22-31.
Williams, S. L., & Rappoport, A. (1983). Cognitive

treatment in the natural environ
ment for agoraphobics. Behavior Therapy, /4,299-313.
Williams, S. L., Turner, S. M., & Peer, D. F. (\ 985).

Guided mastery and performance
desensitization treatments for severe acrophobia. Journal

of Consulting and Clinical
Williams, S. L., & Watson, N. (1985). Perceived danger and

perceived self-efficacy as
cognitive determinants of acrophobic behavior. Behavior

Therapy, /6, 237-247.
Williams, S. L., & Zane, G. (1989). Guided mastery and

stimulus exposure treatments
for severe performance anxiety in agoraphobics. Behaviour

Research and Therapy,
27,237-247.
Wolpe, J. (\ 958). Psychotherapy by reciprocal inhibition.

Stanford, CA: Stanford
University Press. Author Notes The preparation of this

paper and much of the research reported in it was
supported by United States Public Health Service grants
R03-MH41595 and
R29-MH42385. SELF-EFFICACY AND DEPRESSION David J.

Kavanagh There is now substantial support for a
correlation between depression and reduced judgments of
self-efficacy. However, there are several possible paths
to this result: For example, lower self-efficacy may be
making people depressed, the depression may be undermining
their self-efficacy, or depression may be indirectly
affecting self-efficacy through an impact on performance
attainments. The current evidence suggests that all three
effects are probably occurring. A model is presented, where
self-efficacy judgments, performances and moods have
reciprocal influences on each other, and limiting
conditions for the effects are discussed.
"I am totally hopeless," a depressed woman said to me

recently. "I'm no good at
anything. I can't get any good ideas for my projects, I

keep saying dumb things
when I'm around other people. I look at my friend-she has

a great job, she's
popular, she always seems to know what to say. I'm never

going to be like her.
There's no point in trying." An important aspect of this

person's cognition is her low self-efficacy. This
may be a major factor in her depression being maintained,

since it is likely to in
fluence the activities that she chooses to engage in and

the effort and persistence
she invests in them (Bandura, 1982; Kavanagh, 1983). When

she gives up try
ing, it reduces the opportunities for her to experience

positive outcomes, and
increases the frequency of aversive events for her

(Lewinsohn & Libet, 1972). Her low self-efficacy does not
just seem to affect her performance: It also
seems to be affecting her emotional state directly. When

she talks about her per
ceived incapability in areas that she values highly, she
becomes visibly upset.
This does not seem to be solely because she anticipates

negative external out
comes, but because she is violating her own standards. The

impact appears to be
compounded by an unflattering comparison with other people

and a sense that
this is a permanent problem. The phenomenon is modelled

in Figure I. Depressive feelings that are trig
gered by cognitions, aversive events or physiological

states is thought to have a
reciprocal relationship with both self-efficacy and

performance. As with the cli
ent just described, a low sense of self-efficacy often

deepens the person's
sadness, especially when it makes the opportunity for

positive outcomes seem
remote or when the performance domain is crucial to the

person's self-esteem.
When aversive outcomes occur, these also feed into the

depressive mood (the
arrow on the bottom right). Emotional states are expected

to affect self-efficacy
both directly and through an impact on performance. Aspects

of this model are
supported by the current state of the research, although

some of the influences ap
pear to be less strong than others, and some only occur in

restricted circumstances. SELF-EFFICACY EMOTIONAL STATES
Figure] Theorized relationships between self-efficacy,

performance accomplishments and depression. From Yusaf and
Kavanagh (1990). Is Depression Associated With Low
Self-Efficacy? Depressed people often display reductions
in self-efficacy (Cane & Gotlib,
1985; Davis & Yates, 1982; Kanfer & Zeiss, 1983; Miller,
1984). However this
could arise from a number of processes: The depressive

mood might be reducing
self-efficacy, self-efficacy may be producing sadness, or

the effects may be medi
ated by differences in performances. It may even be the

case that something
outside the model is producing the effects. To determine

whether emotional states are influencing self-efficacy we
need
to manipulate the emotion. In Kavanagh and Bower (1985),

students were asked
to visualize three situations under hypnosis. One of these

was neutral in tone
(sitting at home reading a textbook). The other two were

more emotive situa
tions, involving an interaction with a romantic partner. In

one they had commu
nicated well and had a successful interaction, in the other

they had failed
completely and had been rejected. They were asked to

re-create the feelings of
sadness or happiness that they had experienced in the

situation. Then, they were
asked to rate their self-efficacy on a wide range of

activities. Students who had
visualized a positive romantic event had higher average

levels of self-efficacy
than those who recalled a disastrous experience (Figure 2).

This effect did not
only occur on activities that were associated with romantic

relationships, but also
on more general social skills, and on miscellaneous
activities from athletic tasks
to weight reduction or handling snakes. Since the change

in self-efficacy ratings
did not significantly differ across these domains, we

argued that the results were PERFORMANCE ACCOMPLISHMENTS
unlikely to be due to a generalization of self-efficacy

(which would be expected
to produce a generalization gradient). 65 w 60 0: 0 U

(/) >55 u « u u. u. 50 w 45 SAD NEUTRAL HAPPY
MOOD
Figure 2 Efficacy scores averaged across items for judges

who were happy, neutral, or sad. From Kavanagh and Bower
(1985). Later experiments have attempted to confirm that
the effects could be obtain
ed if the emotion induction had no mention of success or

failure (Kavanagh,
1983; Kavanagh & Hausfeld, 1986). The outcome has been

that the effect is
obtained on some tasks, but-at least for emotions that are

induced in the labora
tory-the impact of the emotions is relatively small. For

example, in Kavanagh
and Hausfeld (1986) happiness produced higher self-efficacy

than sadness on one
of the tasks (push-ups), but not on the other (handgrip).

When subjects were
happy they thought they could do 2.6 more push-ups than

when they were sad.
Since differences of this size are liable to be

overshadowed by within-group vari
ation, the phenomenon is more easily observed in repeated

measures on the same
subjects than in between-subject designs (Kavanagh, 1987).

What then is producing the self-efficacy changes, and why
aren't they strong
er? There is some evidence that the changes in

self-efficacy are mediated by dif
ferences in recalled performances under happy and sad

moods. In Wright and
Mischel (1982), sadness or happiness was induced, and then

subjects were given
bogus feedback about performance on a perceptual task. At

the end of the task,
self-efficacy was assessed and subjects were asked to

recall their success over the
session. When the affective tone of the feedback was

consistent with their mood,
the subjects' predictions about future performances were

close to the actual feed
back information they had received. However, a mismatch

between the emotion
and the feedback information skewed the subjects'

predictions. Happy subjects
who had "failed" predicted higher future performance than

their feedback sug
gested, and sad subjects who had "succeeded" thought they

would do worse.
Since the recalled feedback was similarly skewed, the

self-efficacy effect
appeared to be based on differences in memory for the

performance information. The Wright and Mischel study
(1982) did not clarify whether the impact of
the emotions was primarily occurring at encoding or at

retrieval. We know from
other research that substantial effects can be observed on

encoding. People who
are sad or depressed attend to negative aspects of their

experience more than
happy people do (Bower, 1983; Lishman, 1972; Lloyd &

Lishman, 1975). They
also evaluate their own performance more negatively

(Forgas, Bower, & Krantz,
1984; Lobitz & Post, 1979; Smolen, 1978) and reward

themselves less readily
(Gotiib, 1981), although they are not necessarily less

"accurate" in their self
evaluations (Lewinsohn, Mischel, Chaplin, & Barton, 1980;

cf. Dykman, Horo
witz, Abramson, & Usher, 1991). But influences on

self-efficacy that arise from
the encoding of performance information rely on the person

having undertaken
the activity while sad. Sometimes this has not

occurred-for example, the person
may have never felt sad when playing billiards. Or their

performance on this
particular task may not have been poorer than usual at

times when they were sad. Selective retrieval of
mood-consistent information can also occur (e.g., Clark
& Teasdale, 1982; cf. Bower & Mayer, 1985), and current
thinking suggests that
it is mediated by a differential evaluative response to

the judgment target under
happiness or sadness (Bower, 1991; Schwartz & Bless, 1991).

There is still sub
stantial controversy about the nature of this phenomenon.

In the view of
Schwartz and Clore (1988), people often misread their

current affective state as a
response to the target. This predicts that the effects of

current emotions on both
evaluations and retrieval of consistent memories will be
greater when the reason
for the person's affective state is not salient, and there

is some evidence to sup
port that idea (e.g., Schwartz & Clore, 1983). Bower's

network theory predicts
an impact of the emotion whether the reason for the

emotion is salient or not
(Bower, 1981, 1991). Bower (1991) has drawn attention to

the parallels between cognitive disso
nance and effects of emotions on judgments. But the

situation where people
make successive judgments under different moods is

somewhat different from
the traditional dissonance paradigm. If subjects become

aware of the discrepancy
between their judgments over different days, this might

itself create cognitive
dissonance. This new dissonance might only be resolved by

discounting the
mood effects. Such an effect could explain the results of

Schwartz and Clore
(1983). Verbal reports from subjects suggest that this is

precisely what has
occurred in some of my mood induction experiments: A very

common statement
was along the lines, "I felt that I couldn't do the task,
but then I realised that it
was just because I was feeling unhappy just now, so I put

down what I usually
can do." That is, some subjects seem to be aware of the

effect that the emotion
has had on their judgments or recollections, and actively

adjust for it. The example also suggests that these
subjects have access to performance
information that conflicts with their prevailing emotion.

This illustrates the point
that both encoding and retrieval effects from emotions may

be overridden by
other salient perfonnance infonnation (Bandura, 1982). I

vividly recall another
subject's response to a self-efficacy question about

lifting weights. "This ques
tion's easy," he said. "I've just come from the gym and I
know what I can lift
today." It would be very difficult for a change in his mood

to affect that self
efficacy judgment! Sometimes the impact of emotions on

self-efficacy may be produced by
mechanisms that have little to do with recall. In this

case, the different recollec
tions of perfonnance may follow the judgment, rather than

producing it. There
are at least two candidates for the alternative mechanisms.

One is that cognitions
about low generalised capabilities have been rehearsed so

frequently in the nega
tive mood that they have become very highly available in

that mood. For exam
ple, the statement "I am totally hopeless" by the client

at the beginning of the
chapter was made several times during the first interview.

When she made a
judgment about a specific capability, this general

proposition may have imme
diately sprung to mind. This would tend to act as a low

anchor for the self
efficacy judgment. We know from data on numerical anchors
(Cervone & Peake,
1986) that people do not correct sufficiently for changes

in their starting point
when they judge their self-efficacy. As a result, her

self-efficacy when she felt
"totally hopeless" would be lower than when she felt happy

and "reasonably
competent." A final possible source of the reductions in

self-efficacy is that people may
feel unable or unwilling to muster the effort that will be

required to perform the
task. This will be more likely when the task is expected to

require substantial
effort for success (see below). Major depressive episodes

and dysthymia may be especially prone to lower
ed self-efficacy because the conditions for all of these

processes may often be
present. The aggregated duration of the disorder makes it

very likely that a wide
range of activities will have been experienced under a

dysphoric mood. If the
depression has strong endogenous features, its origin will

often be unclear to the
person-the situation where Schwartz and Bless (1991)

predict that judgmental
effects will be most prominent. Depressed people are prone

to rehearse genera
lised negative cognitions about themselves (Beck, 1991),

and their judgments
may well be subject to anchoring effects. Unlike the mood

induction subjects,
they also know that their emotional state will probably

continue for some time,
and as a result it may not seem adaptive for them to adjust

their expectations
upwards. No wonder that they feel unable to do well!

Effects of Depression on Peiformance Low expectations of
depressed people are rendered even more plausible
when they notice reductions in their performance. Poorer

perfonnance is often
observed in depression (e.g., Cohen, Weingartner,

Smallberg, Pickar, & Murphy,
1982; Miller, 1975). Some tasks involve a non-depressed

presentation for opti
mal performance: For example smiling often enters an

assessment of social per
fonnance, and a reduction of smiling by depressed people

may partially explain
their poorer social perfonnance scores (e.g., Lewinsohn et

aI., 1980). In addition,
depressed people are more self-absorbed (Ingram & Smith,

1984), and this de
creases their ability to pay attention to others'

contributions and offer empathic
responses. Perhaps because of the preoccupation with

negative thoughts, depres
sed people also do poorly on complex laboratory tasks that

make high attentional
demands (Cohen et aI., 1982). People who are suffering from

a severe Major
Depressive Episode may also feel lethargic (e.g., because

of sleep disturbance
and a reduced food intake) and physiological effects on

psychomotor speed can
occur: These features have a direct impact on timed

responses. However, most depressive deficits occur in
situations (a) where subjects can
withdraw from the activity, or (b) where effort makes an

impact on the perfonn
ance (Ciminero & Steingarten, 1978; Loeb, Beck, & Diggory,

1971; Miller,
1984). Frequently an apparent deficit can be eliminated by

just encouraging peo
ple to persist in their attempts (e.g., Friedman, 1964).

These reductions in persis
tence and effort resemble deficits that are produced by

self-efficacy (Bandura,
1982). In tenns of the model in Figure I, the route of

influence for these partic
ular perfonnance deficits may be primarily through the

changes in self-efficacy
discussed in the previous section. Kavanagh (1987)

examined this possibility in a mood induction experiment.
Happy and sad moods were induced by using a combination of

appropriate music
and recollections of romantic experiences in which

connotations of success or
failure were minimized. A between-subjects design was used,

and a neutral
mood condition was also included. Strong behavioral

differences resulted from
the mood inductions: Subjects who were happy persisted 50%

longer at solving
anagrams than did those who were sad (Figure 3), and spent
a smaller proportion
of their time on the easiest anagrams. Happy subjects also

solved more ana
grams, but this was mainly due to the women. Sad women not
only worked on
the task for a shorter time, they also took four times
longer on average to solve
each anagram. On the other hand, men responded to both

happiness and sadness
by increasing their efficiency. Despite the occurrence of

at least some of the expected perfonnance effects,
there were no parallel effects on self-efficacy before the

anagrams task. In fact,
rather than demonstrating the mediation of self-efficacy

on the perfonnance
changes, the performance attainments produced alterations

in self-efficacy. After
the task, happy subjects did have higher self-efficacy than

those who were sad.
While later perfonnances may be affected by flow-on effects

from these self
efficacy changes, the initial impact of the emotions on

the anagram performance
seemed to be produced by other mechanisms. What could

these other mechanisms be? One possibility is the
maintenance
of positive mood. In Kavanagh (1987), verbal reports

suggested that some sub
jects tenninated the anagrams task because it was giving an

insufficient rate of
positive outcomes, and they aniticipated more positive

experiences (and hence a
better chance of improving their mood) from other

activities. In the case of
women, there was also an impact on cognitive

effi"ciency-perhaps because of
cognitive intrusions that were triggered by the sad

recollection. Why this did not
occur for men is unclear. Other data suggests that they are
more likely to active
ly correct for mood effects on self-efficacy and

performance (Kavanagh, 1983),
and may become more task-oriented as a way of distracting

themselves from
negative recollections. If so, this may be an important

reason for the lower risk
of depression that is observed in men (Nolen-Hoeksema,

1987). Where the pre
dominant incentives for performance are positive, there may

also be an effect on
performance through a reduction in expected pleasure

during sadness. When
subjects are in a negative mood, they expect less enjoyment

from activities
(Carson & Adams, 1980). Positive incentives therefore have

less weight.
Figure 3 Persistence on an anagrams task under sad, neutral

and happy moods. From Kavanagh (1987). Self-Efficacy and
Skills as Determinants of Depression In the example at the
beginning of the chapter, the client reported feeling
worse when she thought about her lack of capabilities.

This suggests that asso
ciations between depression and self-efficacy may partly be

due to an influence
of self-efficacy on emotion. If so, this has important

implications for the treat
ment and prevention of depressive episodes. Despite

considerable debate over the role of cognitions in
depression (e.g.,
Coyne & Gotlib, 1983), cognitive-behavioral approaches to

depression have a 24 22 20 18 M I N U T E S 16 menu 14 12
10 SAD NEUTRAL HAPPY women
strong influence on current psychological theory and

practice (e.g., Beck, 1991;
Beck, Rush, Shaw, & Emery, 1979; Rehm, 1977). One of the
influential theories
has been learned helplessness theory (Abramson, Seligman, &

Teasdale, 1978;
Seligman, 1975). In the original theory, an independence

of responses and out
comes was seen as a cause of depression. When people

experienced negative
outcomes that were not contingent on their responses, they

displayed cognitive,
behavioral and emotional responses that were similar to

those of depressed peo
ple. However, Bandura (1978) argued that the experience of

non-contingent
negative outcomes only produces despondency when the person

ascribes the
negative outcome to their personal inefficacy. This was

highly consistent with
the results that were emerging from the laboratory research

on learned help
lessness (e.g., Klein, Fencil-Morse, & Seligman, 1976). In

response to the evident inadequacies of the original
theory, Abramson et
al. (1978) advanced a revision distinguishing "universal

helplessness"
-approximately the same as the old helplessness

concept-from "personal help
lessness." In personal helplessness, people expect that

the outcome is contingent
on the responses of relevant others, but is not contingent

on any response in their
own repertoire. Except in cases of personal victimization,

this is approximately
equivalent to the concept of low comparative
self-efficacy: Like the client at the
beginning of the chapter, these people usually see

themselves as being unable to
achieve the outcome solely because of their own inability.

In Abramson et al.
(1978), either type of helplessness was expected to

produce despondency if the
outcome was sufficiently aversive, but greater negative

affect was predicted in
personal helplessness, and an association between internal

attributions and
depression was expected. Bandura has continued to assert

that personal inefficacy is the key to
despondency. In his own words: "When people have a low

sense of personal ef
ficacy and no amount of effort by themselves or

comparative others produces
results, they become apathetic and resigned to a dreary

life. The pattern in which
people perceive themselves as ineffectual but see similar

others enjoying the
benefits of successful effort is apt to give rise to

self-disparagement and
depression. Evident successes of others make it hard to

avoid self-criticism"
(Bandura, 1982, p. 141). So, the distress of the client

at the beginning of the chapter appears to be
exacerbated by the comparison with her friend's

capabilities and achievements
("1 look at my friend-she has a great job, she's popular,

she always seems to
know what to say. I'm never going to be like her.").

Bandura's emphasis on personal inefficacy is supported by
the observation
that depressed and nondepressed people do not differ in

their predictions of how
others will do-only on how well they themselves will

perform a task (e.g.,
Garber & Hollon, 1980). In particular, depressed people

lose the self-enhancing
bias in efficacy judgments that is seen when nondepressed

people make social
comparisons (Ahrens, Zeiss, & Kanfer, 1988). Pessimism

about future outcomes
is also linked to the degree of illusory personal control:

If someone else throws
some dice, depressed people expect a higher degree of

success than when they
throw the dice themselves (Golin, Terrell, Weitz, & Drost,

1979). In contrast,
nondepressed people are more optimistic when their personal

control seems
higher. An emphasis on personal inefficacy is also

supported by a relationship
between proneness to depressive mood and a tendency to

ascribe poor perform
ances to oneself (Seligman, Abrams·on, Semmel, & von

Baeyer, 1979). Further
more, improvements in depressive mood during treatment

appears to be corre
lated with increases in perceived mastery or self-efficacy

(e.g., Teasdale, 1985;
Zeiss, Lewinsohn, & Mufioz, 1979). While these studies

support the focus on personal self-efficacy in depression,
a critical test of self-efficacy as a determinant of

depressive mood would require
an examination of the effects from manipulating
self-efficacy. Studies that have
reduced self-efficacy do demonstrate that a depressive mood

can be triggered,
and the effect is more likely when others appear able to

emit the response (e.g.,
Davis & Yates, 1982). The focus on self-efficacy rather

than just on attributions draws attention to
the fact that depressed people are not only concerned about
the cause or blame
for past events-as important as this can be. They are also
concerned about the
implications of the events for the future (cf. Beck et

al., 1979). This is a similar
point that Abramson et al. (1978) made when they emphasised

attributions to
"stable" factors. If the person expects that potentially

aversive situations may
recur, and that they will once again be unable to prevent a

negative outcome, this
is likely to compound their reactions to a current event.

As a result, depression
levels are expected to be related both to the aversiveness

of past events and to
self-efficacy about control of situations they will be

facing in the future. The
point is illustrated in a study by Cutrona and Troutman

(1986). As any new par
ent will attest, the sound of their crying infant induces

significant emotional
distress in the parents (Frodi, Lamb, Leavitt, & Donovan,

1978). When the baby
has a more stormy temperament, the mother is expected to

be at a greater risk of
postpartum depression, and this was indeed the case in the
Cutrona and Troutman
study. But at least some of the relationship between

infant difficulty and
depression was mediated by changes in parenting

self-efficacy. When the
infant's crying was seen by the mother as showing her lack

of capability as a
parent, the risk of depression was amplified. Low

self-efficacy about future situations can of course
produce a range of
negative emotional responses apart from sadness. For

example, anxiety or fear
seems to occur when people anticipate that they may be

unable to control a
potentially aversive situation (Bandura, 1988). On the

other hand, the focus in
sadness appears to be the anticipated loss of positive or

pleasurable experiences
(Bandura, 1982). The reductions in self-efficacy achieve

special significance
when these expected losses are greater-for example, when

the self-efficacy defi
cit covers a wide range of task domains and the person

sees no chance of future
improvement. Unfortunately global deficits are very likely

in sadness (Kavanagh
& Bower, 1985), and like the woman at the beginning of the
chapter, many de
pressed people say they are "no good at anything" and they
will "never" improve. Not only are the losses predicted
from external sources: Sadness is often
triggered by the person's self-reactions. In this case,

the expected loss involves a
withdrawal of self-valuation. While the self-standards of
depressed people may
not exceed those of other people, they frequently are

below the level they think
they can achieve (Kanfer & Zeiss, 1983). When the

performance domain is cen
tral to the person's self-esteem, awareness of the

discrepancy between the stand
ard and the self-efficacy level may have a powerful

immediate impact on the
emotional state. When we take a view of depression that

emphasises both the aversiveness of
experiences and the self-efficacy and skills that people

have in dealing with the
situation, this opens up exciting possibilities for

predicting the course of depres
sion and for developing better methods to prevent its

recurrence. In considering these issues, I have recently
become interested in the process
of recovery from a bereavement (Kavanagh, 1990). Here is a

situation that gene
rates extremely strong emotional reactions and often

produces major changes to
goals and incentives. Yet even after the death of a

spouse, only about a third of
people develop a major depressive episode (Bomstein,

Clayton, Halikas,
Maurice, & Robins, 1973) and most people report some

relief from dysphoria
within 1-3 months after the bereavement (Parkes, 1970).

This is an extreme
example of the low correspondence that is typically found

between life events
and depression, where events per se account for only about
10% of the variance
in depression (Brown & Harris, 1986). Some of the reasons

for recovery probably include features of the bereave
ment situation such as the difficulty in avoiding extended

stimulus exposure and
resultant habituation. It is also difficult to avoid some

re-engagement in activi
ties. But a substantial contributor to the high recovery

rate seems to be the skills
that survivors already have in dealing with aversive

events-coping skills that
they have developed from their previous encounters with a

variety of situations,
from broken love affairs to failed examinations or job

retrenchment. At least
some of the existing coping skills can be applied to the

bereavement and
moderate its aversive emotional impact. Consistent with

this idea, Bomstein et
a1. (1973) found a lower risk of depression among widows

who had experienced
a previous bereavement than for those who had not (4%,

compared with 25%). The tasks that are posed by a
bereavement seem to fall into four groups. One
group involves attacking cognitive sources of negative

mood (Beck, 1991; Beck
et aI., 1979). This consists of paying attention to

negative cognitions and
examining their evidence base. A second group involves

solving practical
problems and coping with day-to-day demands such as

managing finances, find
ing new employment, or arranging child care (D'Zurilla &

Goldfried, 1971).
Both of these task areas involve thinking about the

bereavement and its effects.
However in the short term these thoughts will invoke

sadness, especially since
many of the negative cognitions are not easily discounted

and the central problem
is not readily resolved ("the person is gone and will never

come back"). There
fore a third set of tasks involves restoring a more

positive mood through attention
diversion and enjoyable activity (Lewinsohn & Libet,

1972). Here, people use
both activities and cognitive strategies such as fantasy or

humour to divert their
attention from depressive thoughts and to invoke positive

emotion. Finally, the
person may need to be mobilising assistance from other

people, to supplement
deficiencies in their performance of the other tasks

(Cohen & Wills, 1985). When the circumstances of the
death make it very difficult to successfully
undertake these tasks, we would expect people to be more

at risk of depression.
This does seem to be so. When the death involves some blame
being attached to
the survivor-such as when a child dies accidentally (Nixon

& Peam, 1977), or a
spouse suicides (Parkes & Weiss, I 983)-there is a severe

challenge to cognitive
restructuring skills, and the survivor is at special risk

of depression. If the death
was violent the grief reaction can last for many years
(Lehman, Wortman, &
Williams, 1987): In this case, the person is often plagued
by vivid and extremely
distressing images of the death that are very difficult to

suppress. Ongoing legal
action offers further complications by reinvoking the

negative emotions and con
tinuing to disrupt activities. While the specific

challenges in bereavement are often different from other
instances of sadness or depression, the same task areas

apply. There is often a set
of problems to be resolved, negative thoughts are

triggered, the person needs to
find ways to restore a positive mood, and assistance from

others may be required.
Consistent with this view, cognitive-behavioral treatments

of depression that
build skills and performance in these task domains have

achieved impressive
results in controlled trials (Wilson, 1989). We can also

predict future occur
rences of depression from reports of past achievements in

handling daily hassles
and life events, even when the effects of current

depression levels are taken into
account (Holahan & Holahan, 1987). Other research

demonstrates that depressed
people who use active coping strategies (increasing

activity levels, engaging in
interpersonal activity, or rehearsing positive cognitions)

have shorter and less
severe depressive episodes than those who ruminate about

their depression
(Morrow & Nolen-Hoeksema, 1990). From studies of

unselected subjects we
also know that most people have intrusive negative thoughts

from time to time
and that they regularly use cognitive restructuring,

problem solving and attention
diversion strategies to deal with them (Edwards &

Dickerson, 1987). Given these data, self-efficacy
judgments about depression-related task
domains have special significance for understanding the

process of recovery from
depression and predicting its later course (Kavanagh &

Wilson, 1989; Yusaf &
Kavanagh, 1990). In Kavanagh and Wilson (1989), depressed

subjects were
treated with cognitive therapy, and their outcomes over a

12-month follow-up
period were observed. Self-efficacy questions asked about

the subjects' expected
self-control of emotion and cognition over the next 12

months, and attempted to
capture both their capabilities in attacking cognitive

sources of depression and in
positive mood induction. The questions were: (a) "How much

time can you
make at least moderately enjoyable?", (b) "How much time

can you have without
any sad, discouraging or unpleasant thoughts," and (c)

"What percentage of the
negative thoughts that pop into your mind can you

effectively challenge?" In
each case, there were ten levels of performance, from 30

minutes each day to 12
hours or more (or from 0 to 100% in the case of the last

question). Confidence
ratings on the three skill areas were averaged to form a
single score of self
efficacy strength for each subject. Changes in these

self-efficacy scores over the
course of treatment accounted for 50% of the variance in

the depression improve
ment. Furthermore, the self-efficacy scores at

post-treatment strongly predicted
the number of three-month periods that the subjects were

in remission (r = .59),
and the prediction remained significant even after the

post-treatment depression
level was taken into account. The results parallel work

that suggests self-efficacy
about control of anxious thoughts is an important

determinant of anxiety reac
tions (Kent, 1987; Kent & Gibbons, 1987; Ozer & Bandura,
1990). They also
join a range of studies in other problem domains that have

attested the import
ance of self-efficacy for predicting sustained behavioral

improvement after treat
ment (e.g., Kavanagh, Pierce, Lo, & Shelley, 1991;

Sitharthan & Kavanagh,
1990). The precise tasks that will be important for the

depression will of course alter
according to the circumstances that the person is facing.

So, for example, self
efficacy for assertion turned out to be a very powerful

predictor of follow-up
status in the sample used by Yusaf and Kavanagh (1990). In

Cutrona and
Troutman's (1986) sample of postpartum women, parenting

self-efficacy appear
ed especially important. Up to now, 1 have not been

examining the role of self-efficacy about mobilis
ing assistance, but this would appear to be a useful area

to begin looking at. We
know that social support can moderate the aversive impact

of life events (e.g.,
Cohen & Wills, 1985) and that it has a function in

reducing the risk of relapse
after recovery from a depressive episode (Belsher &

Costello, 1988). We also
know that people often find it aversive to talk with people

who are depressed
(Coyne, 1976), and that depressed college students offer

negative self-disclosures
in situations that others find inappropriate (Jacobsen &

Anderson, 1982; Kuiper
& McCabe, J 985). While most of the literature has focused

on initial acquaint
ances among college students, it suggests that skills in

effectively eliciting sup
port and in preserving the sources of support for future

crises may tum out to be
an important focus for assessment and intervention.

Summary There is now substantial support for the model
described in Figure 1, and we
now know in more detail what the limiting conditions are

for the effects that are
described. Reductions in self-efficacy and performance

appear to be both conse
quences of depression and determinants of it. The direct

influence of sadness on
self-efficacy is broad in scope but-at least in studies on

induced moods-it is
often small in degree. The impact of depression on
performance occurs primarily
through an impact on task selection, persistence and

effort, although some direct
effects from depressive symptoms may also be observed.

Poorer performances
are most likely to affect later depression (a) when they

mean that a valued extern
al outcome is not obtained (or an aversive situation is not

terminated), (b) when
they induce negative self-reactions, or (c) when the task

involves the control of
negative mood. These are also the conditions in which low

self-efficacy has an
emotional impact. Self-efficacy judgments appear to affect

emotional state both
directly (when people imagine the future consequences of

their inefficacy), and
through their effect on later performance. However, these

indirect effects of the
self-efficacy judgments are better substantiated by studies

on the prediction of
depressive episodes than by laboratory experiments on

induced mood. The work on sadness and self-efficacy has
contributed to the theory and
research on self-efficacy in two ways. First, it has

extended the role that emotion
has on the formation of self-efficacy judgments. We now

know that emotions
not only act as one of the pieces of information that

people use (as in Bandura,
1977), but that within constraints, they have the potential

to color the other infor
mation that is gained from the other sources (Bandura,
1986). The second con
tribution is the extension of self-efficacy theory on the

prediction and causation
of depressive episodes. The current data provide further

evidence of the signifi
cance of self-efficacy for the production and maintenance

of behavioral change,
Abramson, L. Y., Seligman, M. E. P., & Teasdale, J. D.

(1978). Learned helplessness in humans: Critiques and
reformulation. Journal of Abnormal Psychology, 87, 49-74.
Ahrens, A. H., Zeiss, A. M., & Kanfer, R. (1988).

Dysphoric deficits in interpersonal standards,
self-efficacy, and social comparison. Cognitive Therapy

theory of behavioral change. Psychological Review,
84,191-215.
Bandura, A. (1978). Reflections on self-efficacy. Advances

in Behavior Research and Therapy, 1, 237 -269.
Bandura, A. (1982). Self-efficacy mechanism in human

agency. American Psycholo
gist,37,122-147.

action: A social cognitive theory. Englewood Cliffs, NJ:
Prentice-Hall.

Anxiety Research, 1, 77-98.
Beck, A. T. (1991). Cognitive therapy: A 30-year

retrospective. American Psycholo
gist, 46,368-375.
Beck, A. T., Rush, A. 1., Shaw, B. F., & Emery, G. (1979).

Cognitive therapy of
depression. New York: Guilford.

Belsher, G., & Costello, C. G. (1988). Relapse after
recovery from unipolar depression: A critical review.
Psychological Bulletin, 104,84-96.
Bornstein, P. E., Clayton, P. 1., Halikas, 1. A., Maurice,

W. L., & Robins, E. (1973).
The depression of widowhood after thirteen months. British

Journal of Psychiatry,
122,561-566.
Bower, G. H. (1981). Mood and memory. American

Bower, G. H. (1983). Affect and cognition. Philosophical

Transactions of the Royal
Society of London, Series B, 302,387-402.
Bower, G. H. (1991). Mood congruity of social judgments. In

1. P. Forgas (Ed.).
Emotion and social judgments (pp. 31-53). Oxford: Pergamon.
Bower, G. H., & Mayer, 1. D. (1985). Failure to replicate

mood-dependent retrieval.
Bulletin of the Psychonomic Society, 23,39-42.
Brown, G. W., & Harris, T. (1986). Establishing causal

links: The Bedford College
studies of depression. In H. Katschnig (Ed.). Life events

and psychiatric disorders:
Controversial issues (pp. 107-187). Cambridge: Cambridge

University Press.
Cane, D. B., & Gotlib, I. H. (1985). Depression and the

effects of positive and negative
feedback on expectations, evaluations, and performance.

Research, 9,145-160.
Carson, T. P., & Adams, H. E. (1980). Activity valence as a

function of mood change.
Journal of Abnormal Psychology, 89,368-377.
Cervone, D., & Peake, P. K. (1986). Anchoring, efficacy and

action: The influence of
judgmental heuristics on self-efficacy judgments and

behavior. Journal of Personal
ity and Social Psychology, 50, 492-50 I.
Ciminero, A. G .. & Stein garten, K. A. (1978). The

effects of performance standards on
self-evaluation and self-reinforcement in depressed and

nondepressed individuals.
Cognitive Therapy and Research, 2, 179-182.
Clark, D. M., & Teasdale, J. D. (1982). Diurnal variation

in clinical depression and
accessibility of positive and negative experiences.

Journal of Abnormal Psychology,
91,87-95.
Cohen, R. M., Weingartner, H., Smallberg, S. A., Pickar,

D., & Murphy, D. L. (1982).
Effort and cognition in depression. Archives of General

Psychiatry, 39,593-597.
Cohen, S., & Wills, T. A. (1985). Stress, social support

and the buffering hypothesis.
Psychological Bulletin, 98, 310-357.
Coyne, J. D. (1976). Depression and the response of

others. Journal of Abnormal
Psychology,85,186-193.
Coyne, J. c., & Gotlib, I. H. (1983). The role of cognition

in depression: A critical
appraisal. Psychological Bulletin, 94,472-505.

infant temperament, and
parenting self-efficacy: A mediational model of postpartum
depression. Child
Davis, F. W., & Yates, B. T. (1982). Self-efficacy

expectancies versus outcome expect
ancies as determinants of performance deficitis and

depressive affect. Cognitive
Dykman, B. M., Horowitz, L. M., Abramson, L. Y., & Usher,

M. (991). Schematic
and situational determinants of depressed and nondepressed

students' interpretation
of feedback. Journal of Abnormal Psychology, 100, 45-55.
Edwards, S., & Dickerson, M. (1987). On the similarity of

positive and negative intru
sions. Behaviour Research and Therapy, 25, 207-211.
Forgas, J. P., Bower, G. H., & Krantz, S. (1984). The

influence of mood on perceptions
of social interactions. Journal of Experimental Social

Friedman, A. S. (1964). Minimal effect of severe depression

on cognitive functioning.
Journal of Abnormal and Social Psychology, 69, 237-243.
Frodi, A. M., Lamb, M. E., Leavitt, L. A., & Donovan, W.

L. (1978). Fathers' and
mothers' responses to infant smiles and cries. Infant

Behavior and Development, I, 187-198.
Garber, J., & Hollon, S. D. (1980). Universal versus

personal helplessness: Belief in
uncontrollability or incompetence? Journal of Abnormal

D'Zurilla, J. J., & Goldfried, M. R. (1971). Problem

solving and behavior modification.
Journal of Abnormal Psychology, 78,107-126.
Golin, S., TerreII, F., Weitz, 1., & Drost, P. L. (1979).

The illusion of control among
depressed patients. Journal of Abnormal Psychology, 88,

454-457.
Got\ib, I. H. (1981). Self-reinforcement and recall:

Differential deficits in depressed
and nondepressed psychiatric patients. Journal of Abnormal

Psychology, 90, 521
530.
Holahan, C. K., & Holahan, C. J. (1987). Life stress,

hassles, and self-efficacy in aging:
A replication and extension. Journal of Applied Social

Ingram, R. E., & Smith. T. W. (1984). Depression and

internal versus external focus of
attention. Cognitive Therapy and Research, 8, 139-152.
Jacobsen, N. S., & Anderson, E. A. (1982). Interpersonal

skill and depression in col
lege students: An analysis of the timing of

self-disclosures. Behavior Therapy, 13,
271-282.
Kanfer, R., & Zeiss, A. M. (1983). Depression,


Kavanagh, D. J. (1983). Mood and self-efficacy: Influence

of joy and sadness on
efficacy and task selection. (Doctoral dissertation,

Stanford University, 1983).
Dissertation Abstracts International, 44, 2922B.

Kavanagh, D. 1. (1987). Mood, persistence and success.
Australian Journal of
Kavanagh, D. J. (1990). Toward a cognitive-behavioural

intervention for adult grief
reactions. British Journal of Psychiatry, 157, 373-383.
Kavanagh, D. 1., & Bower, G. H. (1985). Mood and

self·efficacy: Impact of joy and
sadness on perceived capabilities. Cognitive Therapy and

Research, 9, 507-525.
Kavanagh, D. 1., & Hausfeld, S. (1986). Physical

performance and self-efficacy under
happy and sad moods. Journal of Sport Psychology, 8, I J

2-123.
Kavanagh, D. 1., Pierce, J. P., Lo, S. K., & Shelley, 1.

(1991). Self-efficacy and social
support as predictors of smoking after a quit attempt.

University of Sydney
(manuscript submitted for publication).
Kavanagh, D. J., & Wilson, P. H. (1989). Prediction of

response to cognitive therapy
for depression. Behaviour Research and Therapy, 27,

333-343.
Kent, G. (1987). Self-efficacious control over reported

physiological, cognitive and
behavioural symptoms of dental anxiety. Behaviour Research

and Therapy, 25,341
347.
Kent, G., & Gibbons, R. (1987). Self-efficacy and the

control of anxious cognitions.
Journal of Behaviour Therapy and Experimental Psychiatry,

18,33-40.
Klein, D. c., Fencil-Morse, E., & Seligman, M. E. P.
(1976). Learned helplessness,
depression, and the attribution of failure. Journal of

Personality and Social Psychol
ogy,33, 508-516.
Kuiper, N. A., & McCabe, S. B. (1985). The appropriateness

of social topics: Effects
of depression and cognitive vulnerability on self and other

judgments. Cognitive
Lehman, D. R., Wortman, C. R., & Williams, A. F. (1987).

Long-term effects of losing
a spouse or child in a motor vehicle crash. Journal of

Lewinsohn, P. M., & Libet, J. (1972). Pleasant events,

activity schedules, and depres
sion. Journal of Abnormal Psychology, 79, 291-295.
Lewinsohn, P. M., Mischel, W., Chaplin, W., & Barton, R.

(1980). Social competence
and depression: The role of illusory self-perceptions.

Journal of Abnormal
Lishman, W. P. (1972;). Selective factors in memory, Part

2: Affective disorders.
Psychological Medicine, 2, 248-253.
Lloyd, G. G., & Lishman, W. R. (1975). Effect of

depression on the speed of recall of
pleasant and unpleasant experiences. Psychological

Medicine, 5, 173-180.
Lobitz, W. c., & Post, R. D. (1979). Parameters of

self-reinforcement and depression.
Journal of Abnormal Psychology, 88, 33-41.
Loeb, A., Beck, A. T., & Diggory, 1. (1971). Differential

effects of success and failure
on depressed and nondepressed patients. Journal of Nervous

and Mental Disease,
152,106-114.
Miller. G. (1984). An assessment of self-efficacy,

performance and cognitive distortion
following success and failure, in moderately depressed and

nondepressed female
college students. Unpublished Honours Project, University

of Sydney.
Miller, W. R. (1975). Psychological deficit in depression.

Psychological Bulletin, 82,
238-260.
Morrow. J., & Nolen-Hoeksema, S. (1990). Effects of

responses to depression on the
remediation of depressive affect. Journal of Personality

and Social Psychology, 58,
519-527.
Nixon, J., & Pearn, J. (1977). Emotional sequelae of

parents and siblings following the
drowning or near-drowning of a child. Australian and New

Zealand Journal of
Psychiatry, I 1,265-268.
Nolen-Hoeksema. S. (1987). Sex differences in depression:

Evidence and theory.
Psychological Bul/etin, /0 I, 259-282.
Ozer, E. M., & Bandura, A. (1990). Mechansims governing

empowerment effects: A
self-efficacy analysis. Journal of Personality and Social
Parkes, C. M. (1970). The first year of bereavement: A

longitudinal study of the reac
tion of London widows to the death of their husbands.

Psychiatry, 33, 444-467.
Parkes, C. M., & Weiss, R. S. (1983). Recoveryfrom

bereavement. New York: Basic
Books.
Rehm, L. P. (1977). A self-control model of depression.

Behavior Therapy, 8, 787-804.
Schwartz, N., & Bless, H. (1991). Happy and mindless, but

sad and smart? The impact
of affective states on analytic reasoning. In 1. P. Forgas

(Ed.). Emotion and social
judgments (pp. 55-71). Oxford: Pergamon.
Schwartz, N., & Clore, G. L. (1983). Mood, misattribution

and judgments of well
being: Informative and directive functions of affective

states. Journal of Personality
and Social Psychology, 45, 5l3-523.
Schwartz, N., & Clore, G. L. (1988). How do I feel about

it? The informative functions
of affective states. In K. Fiedler & J. Forgas (Eds.).

Affect. cognition. and social
behavior (pp. 44-62). Toronto: Hogrefe.
Seligman, M. E. P. (1975). Helplessness: On depression,

development and death. San Francisco: W. H. Freeman.
Seligman, M. E. P., Abramson, L. Y., Semmel, A., & von

Baeyer, C. (1979). Depres
sive attributional style. Journal of Abnormal Psychology,

88, 242-247.
Sitharthan, T .. & Kavanagh, D. 1. (\ 990). Role of
self-efficacy in predicting outcomes
from a programme for controlled drinking. Drug and Alcohol

Dependence, 27, 87-94.
Smolen, R. C. (1978). Expectancies, mood and performance

of depressed and non
depressed psychiatric inpatients on chance and skill tasks.

Journal of Abnormal
Teasdale, J. D. (1985). Psychological treatments for

depression How do they work?
Behaviour Research and Therapy, 23, 157-165.
Wilson, P. H. (1989). Cognitive-behaviour therapy for

depression: Empirical findings
and methodological issues in the evaluation of outcome.

Behaviour Change, 6, 85-95.
Wright, 1., & Mischel, W. (1982). The influence of affect

on cognitive social learning
person varaibles. Journal of Personality and Social

Yusaf, S., & Kavanagh, D. 1. (1990). The role of

self-efficacy in the treatment of
depression. Journal of Cognitive Psychotherapy, 4, 51-70.
Zeiss, A. M., Lewinsohn, P. M., & Munoz, R. F. (1979).

Nonspecific improvement
effects in depression using interpersonal skills training,

pleasant activity schedules, or
cognitive training. Journal of Consulting and Clincial

Psychology, 47,427-439. This page intentionally left blank
SELF-EFFICACY AS A RESOURCE FACTOR IN STRESS APPRAISAL
PROCESSES Matthias Jerusalem and Ralf Schwarzer According
to the cognitive-relational theory of stress, emotions, and
coping, cognitive appraisals are seen as mediating
processes that refer to the stakes a person has in a
stressful encounter and to the coping options. They result
in either challenge, threat, or harmlloss. It is
undetermined, however, how these appraisals are
interrelated over time and whether they can occur
simultaneously. An idealized motivation model has been
established to stimulate research on this issue. The
present experiment has been set up to assess the dynamic
pattern of cognitive appraisals at nine points in time
under stress, defined as continuous failure at demanding
academic tasks. General self-efficacy is considered to
represent a personal resource among other antecedents of
appraisals. Therefore, self-efficacy was used as a
betweengroups factor. Very different patterns of
appraisals emerged for low and high self-efficacious
subjects, indicating that high self-efficacy buffers the
experience of stress, whereas low self-efficacy puts
individuals at risk for a dramatic increase in threat and
loss appraisals.
The present study deals with the prediction of cognitive

appraisal processes by
dispositional antecedents and by stressful conditions. It

is based on the cognitive
relational theory which defines stress as "a particular

relationship between the
person and the environment that is appraised by the person

as taxing or exceeding
his or her resources and endangering his or her well-being"

(Lazarus & Folkman,
1984b, p. 19). Appraisals are determined simultaneously by

percei ving environ
mental demands and personal resources. They can change over

time due to
coping effectiveness, altered requirements, or improvements

in personal abilities. The cognitive-relational theory of
stress emphasizes the continuous, recipro
cal nature of the interaction between the person and the

environment. Since its
first publication (Lazarus, 1966), it has not only been

further developed and
refined, but it has also been expanded recently to a

meta-theoretical concept of
emotion and coping processes (Lazarus, 1991a; Lazarus &

Folkman, 1987). The
present paper deals with an experimental study based partly

on this meta-theory
of emotions and coping. Therefore, it is necessary to

describe briefly some
aspects of this theory that are relevant for the

understanding of the study. Meta-Theoretical
Considerations Within a meta-theoretical system approach
Lazarus and Folkman (1987)
conceive the complex processes of emotion as composed of

causal antecedents,
mediating processes, and effects. Antecedents are person

variables like commit
ments or beliefs on the one hand and environmental

variables, such as demands
or situational constraints, on the other. Mediating

processes refer to cognitive ap
praisals of situational demands and personal coping

options as well as to coping
efforts aimed at more or less problem-focused and

emotion-focused behavior
(Jerusalem & Schwarzer, 1989; Krohne, 1988; Laux & Weber,

1987; Lazarus &
Folkman, 1987; McCrae & Costa, 1986). Stress experiences

and coping results
bring along immediate effects, such as affects or

physiological changes, and
long-term results concerning psychological well-being,

somatic health and social
functioning (Lazarus & Folkman, 1984a, 1984b). There are

three meta-theoretical assumptions: transaction, process,
and
context. It is assumed, first, that emotions occur as a
specific encounter of the
person with the environment and that both exert a

reciprocal influence on each
other; second, that emotions and cognitions are subject to

continuous change; and
third, that the meaning of a transaction is derived from

the underlying context,
i.e., various attributes of a natural setting determine

the actual experience of
emotions and the resulting action tendencies (Lazarus, 1991

a, 1991 b). For obvious reasons, prior research has mostly
neglected these meta
theoretical assumptions in favor of unidirectional,

cross-sectional, and rather
context-free designs. Within methodologically sound

empirical research it is
hardly possible to study complex phenomena such as emotions

and coping with
out constraints. Also, on account of its complexity and

transactional character
leading to interdependencies between the involved

variables, the meta-theoretical
system approach cannot be investigated and empirically

tested as a whole model.
Rather, it represents a heuristic frame that may serve to

formulate and test
hypotheses in selected subareas of the theoretical system

only. Thus, in practical
research one has to compromise with the ideal research

paradigm. Investigators
have often focused on structure instead of on process,

measuring single states or
aggregates of states. In the present study, however,

stress is analyzed and inves
tigated as an active. unfolding process. More precisely,

stress appraisal processes
are predicted by environmental and personal variables as

antecedents, whereas
coping strategies and long-term effects are not considered.

From the meta-theo
retical system perspective, the study concentrates on

stress antecedents and actual
stress as a process, but has its limits with respect to

transaction and context. In view of the research
intention and the variables involved in the present
empirical study, the concept of stress appraisals will be

discussed first. Second,
some environmental and personal antecedents of stress

evaluations are
considered. Finally, we will refer to the phenomena of

mixed appraisal patterns
and their development over time. Stress Appraisals The

cognitive relational theory (Lazarus & Folkman, I 984a,
1987) defines
stress as an encounter in which the demands tax or exceed

the available re
sources. Cognitive appraisals include two component

processes, primary and
secondary appraisals. Primary appraisal refers to the

stakes a person has in a
certain encounter. In primary appraisals, a situation is

perceived as being either
irrelevant, benign-positive or stressful. Those events

classified as stressful can be
further subdivided into the categories of challenge, threat

and harm/loss. A stress-relevant situation is appraised as
challenging when it mobilizes phy
sical and psychological activity and involvement. In the
appraisal of challenge, a
person may see an opportunity to prove herself or himself,

anticipating gain,
mastery or personal growth from the venture. The situation

is experienced as
pleasurable, exciting, and interesting, and the person is

hopeful, eager, and confi
dent to meet the demands. Threat occurs when the

individual perceives being in danger, and it is experi
enced when the person anticipates future harm or loss. Harm

or loss can refer to
physical injuries and pain or to attacks on one's

self-esteem. Although in threat
appraisal future prospects are seen in a negative light,

the individual still seeks
ways to master the situation faced. The individual is

partly restricted in his or her
coping capabilities, striving for a positive outcome of the

situation in order to
gain or to restore his or her well-being. "Rather, threat

is a relational property
concerning the match between perceived coping capabilities

and potentially hurt
ful aspects of the environment" (Bandura, 1991, p. 90). In

the experience of harm/loss, some damage to the person has
already
occurred. Damages can include the injury or loss of valued

persons, important
objects, self-worth or social standing. Instead of

attempting to master the situa
tion, the person surrenders, overwhelmed by feelings of

helplessness. Beck's
cognitive theory of anxiety and depression (Beck & Clark,

\988) is in line with
these assumptions, mentioning threat as the main cognitive

content in anxiety
compared to loss as its counterpart in depression. Primary

appraisals are mirrored by secondary appraisals which refer
to one's
available coping options for dealing with stress, i.e.,

one's perceived resources to
cope with the demands at hand. The individual evaluates his

competence, social
support, and material or other resources in order to

readapt to the circumstances
and to reestablish an equilibrium between person and

environment. In academic
situations mostly the task-specific competence or the

prerequisite knowledge to
cope with the task is of primary importance. There is no

fixed time order for pri
mary and secondary appraisals. The latter may come first.

Moreover, they depend
on each other and often appear at the same time. Instead of

primary and
secondary, the terms "demand appraisal" and "resource

appraisal" might be more
appropriate. Hobfoll (1988, 1989) has expanded the stress

and coping theory with
respect to the conservation of resources as the main human

motive in the struggle
with stressful encounters. Antecedents of Stress

Appraisals Stress appraisals result from perceived
situational demands in relation to per
ceived personal coping resources. Despite this relational

conception one can ima
gine environmental conditions that are more likely to

induce stress than others,
provided the same person is confronted with them. One can

also imagine individ
ual differences in perceived personal resources that make

people more or less
vulnerable to the same environmental requirements. With

respect to the relevance of situational stressors, Lazarus
and Folkman
(1984b) mention formal properties, such as novelty, event

uncertainty, ambiguity
and temporal aspects of the stressing conditions. For

example, demands that are
difficult, ambiguous, unannounced, not preparable, to be

worked on both for a
long time and under time pressure, are more likely to

induce threat perceptions
than easy tasks which can be prepared for thoroughly and

can be solved under
convenient pace and time conditions. Regarding content,

environmental aspects
can be distinguished with respect to the stakes involved by

the kind of a given
situation. For example, threatening social situations

imply interpersonal threat,
the danger of physical injury is perceived as physical

threat, and anticipated fail
ures endangering self-worth indicate ego-threat (McGrath,

1982; Spielberger,
1985). Lazarus (1966) additionally distinguishes between

task-specific stress,
including cognitive demands and other formal task

properties, from failure
induced stress, including evaluation aspects such as social

feedback, valence of
goal, possibilities to fail, or actual failures. By and
large, unfavorable task condi
tions combined with failure-inducing situational cues are

likely to provoke
feelings of distress. With respect to the relevance of

perceived personal resources, Lazarus and
Folkman (l984b) mention commitments and beliefs.

Commitments represent
motivational structures such as personal goals and

intentions that in part deter
mine perceptions of situational stress relevance and the

stakes at hand (Novacek
& Lazarus, 1990). Provided the stakes are really relevant,

beliefs as personal
antecedents of stress appraisals come into play. Beliefs

are convictions and
expectations of being able to meet situational

requirements. With "generalized
beliefs"-as opposed to situation-specific appraisals of

control-dispositional
resource or vulnerability factors are meant, such as locus

of control, general self
efficacy, trait anxiety, or self-esteem (Folkman, 1984;

Hobfoll, 1989; Jerusalem,
I 990a, I 990b; Lazarus & Folkman, 1987). Given a stressful

situation, low
dispositional control expectancies make people vulnerable

to distress, whereas
perceptions of high dispositional competence represent a

positive resource factor. Since the present study is
concerned with achievement tasks to be solved
under ego-threatening conditions, general self-efficacy can

be conceived of as a
personal resource or vulnerability factor (Bandura, 1986,
1989, 1991, 1992;
Jerusalem, 1990a). People who generally trust in their own

capabilities to master
all kinds of environmental demands also tend to interpret

difficult achievement
. tasks as more challenging than threatening. Their

generalized belief of a positive
self-efficacy in this sense serves as a resource factor

that should buffer against
distress experiences furthering "eustress" perceptions

instead. Individuals, how
ever, who are characterized by generally low self-efficacy

expectations are prone
to self-doubts, state anxiety, threat appraisals and

perceptions of coping
deficiencies when confronted with critical achievement

demands. Moreover, pre
vious research on anxiety and self-related cognitions has

demonstrated that gen
eralized beliefs of weak self-efficacy make persons

vulnerable toward distress
experiences because they tend to be permanently worried,

have weak task
specific competence expectancies, interpret physiological

arousal as an indicator
of anxiety, regard achievement feedback as social

evaluations of their personal
value, and feel more responsible for failure than for

success (Bandura, 1992;
Carver & Scheier, 1988; Dweck & Wortman, 1982; Epstein,

1986; Jerusalem,
1990a, 1990b; Ozer & Bandura, 1990; Sarason, 1988;

Schwarzer, 1986;
Schwarzer & Wicklund, 1991; Spielberger, 1972, 1985; Wine,
1980, 1982;
Wood & Bandura, 1989). Like other trait-like person

characteristics, weak gen
eral competence expectancies have numerous causes. A

history of success and
failure combined with a lack of supportive feedback and an

unfavorable attribu
tional style by parents, teachers and peers may lead to the

development of a tend
ency to scan the environment for potential dangers

("sensitizing"), to appraise
demands as threatening, and to cope with problems in a

maladaptive way. In the
present context it can be summarized that general

self-efficacy is seen as a per
sonal resource factor with respect to distress experiences

such as threat and loss
perceptions, which are assumed to come up faster and to a

higher degree for low
compared to high self-efficacious subjects. The latter are

expected to feel more
challenged instead. These assumptions do not apply to the

absolute level of each
distinct stress appraisal, but rather to the prediction of

appraisal patterns and their
changes over time. For a better understanding of this

phenomenon, the issue of
appraisal patterns and processes is to be discussed in more

detail. Appraisal Patterns and Processes The experience
of challenge, threat or loss does not happen exclusively,
but
can overlap or even occur simultaneously. Challenge,

threat and loss are not to be
considered as clearly distinct modes of experience, but
rather as interrelated cog
nitive-emotional states that exist simultaneously. For

example, a person might be
challenged by the demanding characteristics of a situation,

while expecting injury
to his or her well-being at the same time. Since life often

confronts us with situa
tions that are unforeseen, difficult, novel or ambiguous,

the corresponding ap
praisals can be complex states with more or less favorable

and unfavorable
evaluations of these environmental demands. Moreover, the

structure of this mix
ture is likely to change over time unless the problem is

solved immediately. Once
a transactional stress process has commenced, the pattern

of positive and nega
tive evaluations changes from one encounter to the next.

Challenge appraisals
might be stronger than threat at one time, both could be

perceived as equal size
the next time, and another time threat might exceed

challenge. The actual pattern
at each point in time reflects the momentary subjective

uncertainty of being able
to cope with the demands at hand. In a study by Folkman and

Lazarus (1985), for
example, 94% of a student sample facing an examination

experienced feelings of
both threat and challenge two days before the exam.

Afterwards these appraisals
diminished; in case of poor grades they were replaced by

loss perceptions. Simi
lar patterns were found for emotions connected to cognitive

appraisals (Folkman
& Lazarus, 1985; Gall & Evans, 1987; Smith & Ellsworth,
1985, 1987). According to these research findings,
stressful encounters are dynamic, un
folding processes that imply complex appraisal patterns,

rather than static, uni
tary events. More scientific knowledge is required about

the nature of the inter
relationships among challenge, threat, and loss over time

under certain precisely
defined environmental conditions in general, and about the

role of individual dif
ferences in particular. The following theoretical

considerations lead to an idea
lized model of the potential development and change of

appraisal processes under
certain environmental circumstances and for specific

individual differences. A Process Model We have
developed a process model of cognitive appraisals
(Schwarzer,
Jerusalem, & Stiksrud, 1984; Jerusalem, 1990a) which

extends Lazarus' original
stress theory by integrating ideas from Seligman's

helplessness theory
(Abramson, Seligman, & Teasdale, 1978; Brown & Siegel,

1988; Seligman,
1975). A state of helplessness is predicted as a long-term

consequence of cumu
lative experience of personal uncontrollability.

Accordingly, the theoretical
model of appraisal processes was built for the special case

of continuous failures.
Its purpose was to describe the potential development of
loss of control and per
sonal helplessness by means of cognitive appraisal

processes in academic failure
situations. We argue that at almost any point in time, all

three cognitive apprais
als may occur simultaneously, but to differing degrees,

therefore leading to dif
ferent emotions: Challenge causes curiosity, exploration

and productive arousal
threat causes anxiety and loss of control causes

helplessness or even depression. Looking at the process
model in Figure I, the x-axis represents the number of
failures. The curves characterize the idealized potential

development and change
of challenge, threat and loss perceptions. With continued

experience of failures,
challenge diminishes, while loss enhances from one point in

time to the next.
Threat perceptions first increase and then decrease.

Unexpected failures might be
interpreted as challenges to one's competence, whereas

threat or even loss are
less relevant. If this happens repeatedly, the person will

begin to feel more threat201 ened than challenged, but
will stilJ persist with the task. The highest degree of
threat is located at the point where complete subjective
uncertainty about the next outcome prevails. Later, when
subsequent failure is expected with higher certainty, the
individual will experience loss of control while feeling
less threatened because the loss becomes certain. According
to this model, four idealized motivational stages can be
distinguished: 1. The Challenge Stage is a kind of
"reactance stage." Although the person is challenged by
one or more failures, she retains confidence in her ability
to cope with the demands. High self-efficacy may be
combined with productive arousal, i.e., the tendency to
explore the nature of the task. 2. The First Threat Stage
occurs when failures mount and threat surpasses chalJenge
appraisal anxiety is the dominant emotion. The combination
of anxiety with productive arousal here can be called
"facilitating anxiety" because the person is still
self-confident and persists with the task. 3. The Second
Threat Stage occurs at the culminating point, when there is
complete uncertainty about the next outcome. The threat
appraisal is combined with less challenge and some loss of
control. This mode can be called "debilitating anxiety
state" because intrusive self-related cognitions distract
from the task. The person worries about his or her
performance, capability, and the potential for further
failures. 4. Finally, there is a Loss-oJ-Control Stage.
Loss of control is dominant, replacing the appraisal of
threat. The student becomes helpless and disengages; the
next failure is almost certain. Cognitive appraisals cannot
be predicted by the situation at hand exclusively, for
example by task-specificJailure experiences. Development
of Primary Appraisals Challenge Threat Loss Failure
Experiences Figure I Idealized process model of the
development of cognitive appraisals under continuous
failure. The model is restricted to the special case of
continuous failures, a condition
which is rather exceptional than commonplace under

real-life circumstances. Be
sides, human beings confronted with achievement demands

differ with respect to
their acquired competencies and their generalized beliefs.

Thus, there are also
subject-matter-specific Jactors producing differences with

respect to the propor
tional amounts of challenge, threat and loss. Life-long

learning experiences
establish subjective perceptions of individual resources in

the sense of more or
less stable personality traits. For example, anxious

individuals or those with un
favorable generalized expectancies of self-efficacy often

have a learning history
which includes a number of individual failures or

perceived threats in a variety of
more or less taxing situations. High trait anxiety or low
general self-efficacy
make one vulnerable to demands which endanger self-worth.

When confronting stressful academic demands, individuals
with high general
self-efficacy expectancies or low trait anxiety will most

likely start at the left side
of the model curve. Positive self-evaluation retains the

confidence in the ability
to cope with the demands. This is the reactance stage. Less

self-efficacious
people, however, are more likely to display state anxiety

and develop only little,
if any, positive competence expectancies. When failures

mount and the task is
perceived as becoming more difficult, incremental threat

instead of challenge
would be experienced but the person can still be

self-reliant and persistent with
the task. This person may arrive at the first threat

stage. Supposing she suffers
from her generalized belief of low self-efficacy to such a

degree that her expec
tancy for failure is stronger than that for success, she

would start the sequence at
a point where more loss than challenge is experienced. She

then worries about
performance, has self-doubts and is not at all confident

about her competence.
This refers to the second threat stage. For depressed

individuals or those with
high helplessness scores, a dominance of loss perceptions

can be expected. Only
few failures suffice for them to give up further efforts

because the next failure is
almost certain. This is equivalent to the loss-oj-control

stage. Similar stages have
been proposed by Wortman and Brehm (1975) and by

Heckhausen (1991). Empirical evidence was obtained in a
study with high-school students
(Schwarzer et aI., 1984). Within a two-year period,

different student subpopula
tions became more or less anxious and more or less helpless

over time. As pre
dicted, the anxiety level of some subjects declined, but

only at the expense of
helplessness. Particularly long-term low achievers

developed a tendency to per
ceive less threat and more loss of control in face of

academic demands. Regard
ing the model of appraisal processes, this was a

preliminary pilot field study. The
present research questions are how far similar processes

can be observed within a
laboratory experiment and whether general self-efficacy as

a resource factor
serves to predict the hypothesized appraisal patterns and

processes. Only some
few longitudinal field studies have addressed cognitive

appraisal processes (e.g.,
Covington, Omelich, & Schwarzer, 1986; Folkman & Lazarus,

1985; Jerusalem,
1990a; Schwarzer et a1., 1984). The strength of these

studies was that they tried
to capture the dynamic character of the theory by measuring

cognitions in
naturalistic settings at different points in time, but the

joint consideration of
individual differences and experimental factors is still
lacking. Research Question In the present study, the
research question is directed at appraisal processes
depending on experimental treatment and personality traits.

Stress appraisals are
linked to one personal resource factor (general

self-efficacy) and to one environ
mental demands factor (difficult academic tasks). When the

tasks become more
complex and failures mount, all subjects should lower their

initial level of chal
lenge appraisal and experience more threat and more loss

of control until a per
son's maximal level is attained. It is also expected that

low general self-efficacy
represents a major vulnerability condition which

predisposes the corresponding
individuals to appraise the situation as little

challenging, but more threatening,
and-later-as uncontrollable. Thus, two independent factors,

general self
efficacy and demanding experimental tasks are seen as

influential for the devel
opment of three cognitive appraisals over nine points in

time. Further research
questions are the following: Can overlapping appraisal

patterns be observed, i.e.,
do different cognitive appraisals tum up at the same time?

Do cognitive apprais
als change during a stressful encounter? How do appraisals

of challenge, threat,
and loss develop over time in face of continuous failure?

Is there a differential
development with respect to dispositional self-efficacy
levels? How is continuous
failure experienced by low self-efficacious subjects?

METHOD Sample and Procedure Subjects were 210 adults (108
females and 102 males between 21 and 52
years, with a mean age of 29.8 years) who responded to

advertisements in city
magazines and newspapers. The design of the study is

depicted in Figure 2. Sub
jects filled in a series of various personality scales,

among them a general self
efficacy scale. They were then confronted with nine series,

each consisting of
difficult performance tasks, performance feedback, and

self-report items. The experimental setting contained
task-specific stressors (ambiguity, time
pressure, difficulty, etc.) and ego-threatening failure

conditions. Each of the first
six task sequences was composed of 15 anagrams presented

on a computer
screen. Within each set of anagrams the degree of

difficulty increased from the
first to the fifteenth, and all but the last ones were
solvable. Three further task
sequences, consisting of intelligence test items, were

given by the experimenter
in a paper-and-pencil version. All tasks were described as

cognitive problem
solving, to be performed under time pressure, designed to

measure the more
academic and the more practical aspects of intellectual

ability, respectively. Each
anagram was presented individually on the computer monitor

with a number
appearing under each letter. Subjects were to type a new
sequence of numbers Personal Resources/ Vulnerabilities
Se~·Eflicacy Experimental Tasks Feedback Stress
Appraisals Anagrams Success Challenge Threat
Intelligence Tasks Failure Loss 9 Sequences: 6 Anagrams
and 3 Intelligence Tasks
Figure 2 Design of the experiment.
according to the correct order of the letters fonning the

word that was looked for.
A maximum of 40 seconds was al10wed for each anagram.

Subjects were
instructed to call the next item by pressing a certain key

if they found the solution
before the time was up. Otherwise, a tone was sounded and
the next anagram ap
peared automatically. The introductory information on how

to work with the
computer was provided by the computer program.

Paper-and-pencil intelligence
items were given under comparable difficulty and time

conditions. After each task episode, one group
consistently received fictitious success
feedback and the other group fictitious failure feedback.

The feedback was partly
related to actual performance by a programming device so

the outcome seemed
more credible. The feedback referred both to individual

performance (number of
points achieved) and to the performance within their age

group (social compari
son information). In the success group, subjects were told

that the number of
points achieved was above average; in the failure group

individual performance
was reported as being below average. Achievement feedback

was followed by 21
items addressing various aspects of self-related

cognitions, among them
perceptions of challenge, threat and loss. This cycle

(experimental tasks, feedback, stress appraisals) was
repeated nine
times, six with anagrams and three with intelligence items.

Afterwards, subjects
were paid and debriefed about the manipulated feedback. The

whole procedure
lasted about three hours. The present analysis deals with

the failure condition and
with general self-efficacy as a resource/vulnerability

factor for individual
appraisal processes. The failure group consisted of half

the sample (105 subjects). Instruments General
self-efficacy was measured by a scale we developed with
respect to
the self-efficacy theory of Bandura (1977, 1986; Jerusalem

& Schwarzer, 1986).
In this study a six-item version was used (e.g., "When

facing difficulties, I can
. always trust my abilities."). Its internal consistency

was alpha = .82. In order to
assess cognitive appraisals as dependent variables, we

developed psychometric
scales which had to be short because they were presented

nine times; appraisals
should not take more time and should not be emphasized more
than the problem
solving itself. The Challenge scale consisted of four

items, such as "1 am already
curious about how I will manage the next tasks" (alpha =

.78); the Threat scale
contained three items, such as "I am afraid of not being
equal to the next tasks"
(alpha = .8]); and the Loss scale was represented by four

items, such as "I feel
discouraged and depressed now" (alpha = .83). The internal

consistencies within
each appraisal category were averaged over time. The

response format was a
four-point-scale ranging from not at all to a great deal.

RESULTS The failure group was subdivided by use of median
split of the self-efficacy
scores. Scores below average are conceived of as weak

resources indicating high
vulnerability (n = 46), those above average as strong

resources indicating low
vulnerability accordingly (n = 59). By use of 2 x 9

analyses of variance with re
peated measurements, the self-efficacy factor then served

as one predictor of the
intensity and change of primary appraisals over time. At

any point in time, the three cognitive appraisals occurred
simultaneously.
Throughout the nine sequences, all appraisals remained

present, but to different
degrees. In short, as predicted, there were patterns of

appraisals to be observed
instead of mutually exclusive states of challenge, threat

or loss perceptions. First, the overall results are given
separately for the experimental and the
personality factors. Analyses of repeated measurements

revealed that all apprais
als were changing over time. The corresponding F values

were computed accord
ing to the correction procedure proposed by Geisser and

Greenhouse (1958). Due
to continuous failures, challenge appraisals declined
(FI8, 8241 = 20.32,
P < .001), whereas threat and loss perceptions increased

over time (FI8, 8241 =
1.98, p < .05 for threat and Fr8, 8241 = 5.45, p < .001 for
loss). The nature of all
changes can be characterized by linear trend components

(Challenge: F[ I, 1031 =
78, p < .001; Threat: FrI, 103J = 6.7,p < .05; Loss: F[ I,
1031 = 18.8, p < .001). As expected, general
self-efficacy turned out to be an important predictor of
subjective perceptions following failure experiences. With

respect to all appraisal
qualities, low self-efficacious subjects reported less

favorable stress cognitions
than high self-efficacious subjects. Low self-efficacy was

accompanied by con
siderably lower challenge and higher threat and loss

evaluations (challenge: F[ 1,
1031 = 17.5, p < .001); threat: F[ I, 1031= 18.5, p < .001;

loss: F[ I, 1031 = 14.9,
p < .001). There were even additional Self-Efficacy x Time

interactions predict
ing both threat and loss appraisals and indicating a

stronger increase of these
unfavorable distress experiences for low self-efficacy

compared to high self
efficacy subjects (threat: FI8, 824] == 4, p < .001; loss:

FI8, 8241 == 3.1, P < .01 ).
The observed interaction effects are mainly due to linear

trend differences (threat:
F[ I, 1031 == 6.7, p < .05; loss: F[ 1, 1031 == 9.8, p <

.05). In sum, both environ
mental failure conditions and individual self-efficacy
differences were powerful
antecedents of the unfolding process of stress appraisals.

The dominance structure of primary appraisals was affected
by the two stress
factors, too. At the beginning, different structures were

observed for both self
efficacy groups. High self-efficacy individuals felt more

challenged than threat
ened, and they perceived more threat than loss, as

predicted by the theoretical
model. This was similar for low self-efficacy subjects, but

the difference between
challenge on the one hand and threat or loss on the other

hand was to a lesser ex
tent. Compared with the high self-efficacy group, low

self-efficacy subjects
reported less challenge and more threat and loss-at the

very beginning as well
as overall (when summed up over all points in time). The

results for high self-efficacious persons are presented in
Figure 3, those
for low self-efficacious persons in Figure 4. The

corresponding means and stand
ard deviations are reflected in Tables I and 2,

respectively. Scores in challenge,
threat and loss were each divided by the respective number

of scale items in
order to attain the same unit of measurement.
Table I
Means and Standard Deviations of Challenge. Threat and

Lossfor High Self-Efficacy
and Failure Experiences (n = 59) Challenge Threat Loss

Points in Time M SD M SD M SD 1 2.96 .56 1.85 .65
1.66 .57 2 2.74 .66 2.00 .71 1.70 .62 3 2.62 .69
2.01 .78 1.80 .70 4 2.61 .72 2.02 .79 1.72 .69 5 2.49
.73 2.00 .78 1.77 .72 6 2.50 .78 2.00 .84 1.77 .75
7 2.67 .69 1.84 .68 1.73 .74 8 2.58 .66 1.83 .74 1.71
.67 9 2.57 .69 1.84 .73 1.82 .74
Table 2
Means and Standard Deviations of Challenge. Threat and Loss

for Low Self-Efficacy
and Failure Experiences (n = 46) Challenge Threat Loss

Points in Time M SD M SD M SD 2.54 .47 2.25 .65
1.98 .50 2 2.32 .60 2.36 .67 2.05 .62 3 2.19 .63 2.38
.70 2.05 .74 4 2.14 .69 2.40 .73 2.17 .82 5 2.05 .67
2.40 .72 2.26 .84 6 1.95 .76 2.55 .77 2.32 .78 7 2.15
.74 2.50 .63 2.33 .82 8 2.04 .64 2.55 .70 2.35 .78 9
2.02 .68 2.60 .71 2.39 .81 With respect to changes over
time (linear trend components), both groups can
clearly be differentiated. For high self-efficacy subjects,

challenge perceptions
decreased with continued failure experiences (F[ I ,58) =

15.2, p < .00 I). How
ever, threat and loss did not change significantly but

remained on a low and con
stant level, although failures were repeatedly reported.

Throughout the entire
sequence, challenge cognitions exceeded threat and loss

appraisals. In contrast, low self-efficacy subjects were
much more affected by failure.
Challenge perceptions declined (F[ I ,451 = 34.5, P < .00

I), whereas threat and
loss perceptions increased over time (threat: F[ 1,45) =

9.8, P < .01; loss:
F[ 1,45 J = 18, p < .00 1). The strength of these temporal

trends is slightly weaken
ed by motions that take place after measurement point six:

Contrary to the trend
an increase in challenge perceptions and a similar, but

comparably smaller, de
crease in threat evaluations was observed. These effects
were probably due to the
fact that at this point the task material changed from

anagrams to intelligence
items. In face of the altered demands, both reappraisals

of coping options and
corresponding primary reappraisals occurred. Accordingly,

subjective evalua
tions were not generalized from anagrams to intelligence

tasks, at least not for
challenge. However, these "recovery effects" were limited

to a short time period.
Due to the fact that failure feedbacks did not

dissipate-although the task quality
was different-appraisals again became less favorable.

Looking at the results in terms of temporal patterns, it
is obvious that the
appraisal pattern remained stable for high self-efficacy

subjects in the sense that
challenge exceeded threat and loss perceptions at all nine

points in time. In com
parison, the results for low self-efficacy subjects were

completely different. The
temporal dominance structure of stress appraisals in this

group was reversed: At 2
Figure 3 Temporal pattern of cognitive appraisals for high

self-efficacy subjects.
Figure 4 Temporal pattern of cognitive appraisals for low

self-efficacy subjects.
first challenge clearly dominated threat and loss, but at

the end threat and loss
had become the prevailing appraisals. Accordingly, two

intersection points were
observed: Challenge was surpassed by threat at Point 2 and
by loss at Point 4.
Referring to the theoretical model, these intersection

points indicate shifts from
the reactance stage (Stage 1) to the first threat stage

(Stage 2), and from the first
to the second threat stage (Stage 3), respectively.

However, loss never exceeded
threat perceptions, which would have been required to

indicate a state of
helplessness (Stage 4). Temporal Pattern of Primary

Appraisals Challenge Threat Loss 1 2 3 4 S 6 7 8 9 High
Self-Efficacy / Failure Condition Tempora l Pattern of
Primary Appraisals Challenge Threat Loss 1 2 3 4 5 6 7 8 9
Low Self-Efficacy / Failure Condition DISCUSSION The
reported findings demonstrate mainly that situation
variables and person
variables are prominent predictors of perceived challenge,

threat and loss in
stressful situations. In an experimental setting these

appraisals differed in
dependence on evaluative feedback and general

self-efficacy. In face of con
tinuous failures stress appraisals changed from favorable

to unfavorable evalua
tions in the long run. At the same time, low

self-efficacious individuals felt more
distressed than high self-efficacious individuals. As

predicted, stress appraisals
did not come up as single states that are switched either

"on" or "off." Instead,
dynamic patterns of appraisals do exist that are related

to each other and that are
also changeable according to the transactional processes

involved. By this, the
entire pattern is continuously changing, but at no time
will any of the appraisals
disappear unless the situation is no longer perceived as

stressful. With respect to these complex and dynamic
appraisal processes the role of
general self-efficacy as a resource/vulnerability factor

could be clearly demon
strated. Moreover, the results could be mostly explained

with the help of a theo
retical model. In general, high self-efficacious

individuals seem to be scarcely
vulnerable to the stressing situational conditions since

they hardly show any re
actions to experimental manipulations. Obviously, high

general self-efficacy is
equivalent to or leads to positive personal beliefs

providing people with good
capacities to resist stress. Even though failures are

consistently given at nine
points in time, appraisals remain predominantly positive,

and their relative
strengths do not change. During the whole time sequence,

subjects remain within
the challenge or reactance stage, and they do not enter

threat or loss stages at all.
The stressing situational conditions do not take much

effect because they are con
fronted with a personality who can resist on account of

his or her powerful
resources and low vulnerability, respectively. By and

large, the experimental situ
ation faced by low vulnerability persons seems not to be

stressful enough to
induce strong negative emotional experiences. Contrary to

this obvious stress resistence, low self-efficacious
subjects seem
to be especially vulnerable for difficult demand

characteristics and failure experi
ences. They come up with less favorable evaluations,

consistently developing to a
more negative way of appraising situational demands. By

and by, they shift from
reactance to the frrst threat stage, and later on to the

second threat stage as well.
In this case, the experimental demands and failure

feedbacks do take much effect
because they are perceived by a personality who is

handicapped by his or her
weak resources and high vulnerability, respectively.

However, the experimental
situation has its constraints in so far as it is not

powerful enough to make subjects
turn to stage four in the process model, that is, to

elicit predominant loss-of
control evaluations and overwhelming feelings of

helplessness. However, for
ethical reasons one should not complain about this specific

result. Concerning the theoretical model, the predicted
slope of threat appraisals that
could not be fitted exactly in this investigation might be

particularly sensitive to
the stakes given. Regarding the above-mentioned field study

(Schwarzer et al.,
1984), long-term low grades endanger the academic career

of the respective
students. Since this concerns an important real-life stake

almost everyone strives
for, the low graders have enough reason to feel helpless

and depressed in the long
run. Within an experimental situation one has to undertake

and then is able to
leave again, it is more likely that real-life stakes are

not so deeply involved, i.e.,
central values and commitments might be less emphasized.

In order to feel
severely helpless, individuals most likely have to suffer

from failures in more
stress-relevant real-life situations than only from

negative achievement feedback
in a laboratory setting where ethical limits are given in

this respect. Appraisal processes also turned out to be
sensitive to variations in task mater
ial or problem structure at hand. Movements of challenge

and threat contrary to
the trend appeared when a shift from anagrams to

intelligence problems occured.
One explanation would be that by this shift several demand

characteristics are
changing, too, for example situational ambiguity,

familiarity with the tasks at
hand, or their perceived validity for assessing

intellectual ability. The distinc
tiveness of task material may cause new hope to be better

able to meet the
requirements of the upcoming problems. However, this

renewed confidence
disappears again as soon as further failures are faced. It

is interesting to note that predominantly low
self-efficacious individuals
show strong reactions to task shifts in this way, whereas

high self-efficacious
subjects seem to be less sensitive to differing task

materials. For the latter, their
generalized beliefs obviously provide a shelter from all

kinds of failure feedback,
at least from those experienced in the experimental

session. Maybe they are so
confident in their abilities so as not to attach any

importance to the artificial prob
lems given by the experimenter because they do not believe

in the tasks' validity
for intelligence measurement and/or the ecological validity

of the experimental
condition. In contrast, low self-efficacious individuals

do not possess such a
strong shelter from external feedback because they are much

less secure with
respect to their actual abilities. Due to their lack of

self-confidence they are
guided to a large extent by situational cues which have a

strong impact on stress
appraisals, be it failure, task material change, or other

conditions. General self-efficacy as a
resource/vulnerability factor towards feeling
distressed when confronted with achievement demands serves

as a moderator for
the impact of these demands on actual stress experience as

represented by
cognitive appraisals. High self-efficacy subjects rather

select positive cues, such
as stable and favorable self-evaluations, than negative

signals, such as situational
failures, in appraising the stressing person-environment

relationship. Low self
efficacy subjects see themselves confronted with negative

achievement cues and
negative self-evaluations. Thus, both information sources
are combined and
attached to high importance, leading to distressful

experiences which tum from
Abramson, L. Y., Seligman, M. E. P., & Teasdale, J. (1978).

Learned helplessness in
humans: Critique and reformulation. Journal of Abnormal

Bandura, A. (1977). Self-efficacy: Toward a unifying theory

of behavioral change. Psy
. chological Review, 84,191-215.

Englewood Cliffs, N.J.: Prentice-Hall.
Bandura, A. (1989). Human agency in social cognitive

theory. American Psychologist,
44,1175-1184.
Bandura, A. (1991). Self-efficacy conception of anxiety. In

R. Schwarzer & R. A.
Wicklund (Eds.), Anxiety and self-focused attention (pp.

89-110). London: Harwood
Academic Publishers.
Bandura, A. (1992). Exercise of personal agency through

the self-efficacy mechanism.
In R. Schwarzer (Ed.), Self-efficacy: Thought control of

action. New York:
Hemisphere (this volume).
Beck, A. T., & Clark, D. A. (1988). Anxiety and depression:

An information processing
perspective. Anxiety Research, 1,23-36.
Brown, J. D., & Siegel, 1. M. (1988). Attributions for

negative life events and depres
sion: The role of perceived control. Journal of

55,316-322.
Carver, C. S., & Scheier, M. F. (1988). A control-process

perspective on anxiety.
Anxiety Research, I, 17-22.
Covington, M., Omelich, C., & Schwarzer, R. (1986).

Anxiety, aspirations and seIf
concept in the achievement process. A longitudinal model

with latent variables.
Motivation and Emotion, 10, 71-88.
Dweck, C. S., & Wortman, C. B. (1982). Learned

helplessness, anxiety and achieve
ment motivation. In H. W. Krohne & L. Laux (Eds.),

Achievement, stress, and
anxiety (pp. 93-125). Washington, DC: Hemisphere.
Epstein, S. (1986). Anxiety, arousal, and the self-concept.

In C. D. Spielberger & 1. G.
Sarason (Eds.), Stress and anxiety. A sourcebook of theory

and research (pp. 265
305). Washington, DC: Hemisphere.
Folkman, S. (1984). Personal control and stress and coping

process: A theoretical analy
sis. Journal of Personality and Social Psychology, 46,

839-852.
Folkman, S., & Lazarus, R. S. (1985). If it changes it

must be a process: Study of emo
tion and coping during three stages of a college

examination. Journal of Personality

Gall, T. L., & Evans, D. R. (1987). The dimensionality of
cognitive appraisal and its
relationship to physical and psychological well-being. The

Journal of Psychology,
121,539-546.
Geisser, S., & Greenhouse, S.W. (1958). An extension of

Box's results on the use of the
F-distribution in multivariate analysis. Annals of

Mathematical Statistics, 29, 885
89l.
Heckhausen, H. (1991). Motivation and action. Heidelberg:

Springer.
Hobfo\I, S. E. (1988). The ecology of stress. Washington,

DC: Hemisphere.
Hobfo\I, S. E. (1989). Conservation of resources: A new

attempt at conceptualizing
stress. American Psychologist, 44 (3), 513-524.
Jerusalem, M. (1990a). Personliche Ressourcen,

Vulnerabilitiit und Strej3erleben
[Personal resources, vulnerability, and stress experiencej.

Gottingen: Hogrefe.
Jerusalem, M. (1990b). Temporal patterns of stress

appraisals for highand low
anxious individuals. Anxiety Research, 3, 113-129.
Jerusalem, M., & Schwarzer, R. (1986). Selbstwirksamkeit

[Self-efficacy]. In
R. Schwarzer (Ed.), Skalen zur Bejindlichkeit und

Personlichkeit (pp. 15-28). Berlin:
Freie Universitlit, Institut flir Psychologie.
Jerusalem, M., & Schwarzer, R. (1989). Anxiety and

self-concept as antecedents of
stress and coping: A longitudinal study with German and

Turkish adolescents.
Personality and Individual Differences, 10 (7), 785-792.
Krohne, H. W. (1988). Coping research: Current theoretical

and methodological devel
opments. The German Journal of Psychology, 12, 1-30.
Laux, L., & Weber, H. (1987). Person-centred coping

research. European Journal of
Personality, I, 193-214.
Lazarus, R. S. (1966). Psychological stress and the coping

process. New York:
McGraw-HilI.
Lazarus, R. S. (1991a). Emotion and adaptation. New York:

Lazarus, R. S. (1991 b). Cognition and motivation in

emotion. American Psychologist,
46,352-367.
Lazarus, R. S., & Folkman, S. (1984a). Coping and

adaptation. In D. Gentry (Ed.), The
handbook of behavioral medicine (pp. 282-325). New York:

Guilford.
Lazarus, R. S., & Folkman, S. (1984b). Stress. appraisal.

Springer.
Lazarus, R. S., & Folkman, S. (1987). Transactional theory

and research on emotions
and coping. European Journal of Personality, I, 141-170.
McCrae, R. R., & Costa, R. T. (1986). Personality, coping,

and coping effectiveness in
an adult sample. Journal of Personality, 54, 385-405.
McGrath, J. E. (1982). Methodological problems in research

on stress. In H. W. Krohne
& L. Laux (Eds.), Achievement, stress. and anxiety (pp.
19-50). Washington, DC:
Hemisphere.
Novacek, 1., & Lazarus, R. S. (1990). The structure of

personality commitments.
Journal of Personality, 58,693-715.
Ozer, E., & Bandura, A. (1990). Mechanisms governing

empowerment effects: A self
efficacy analysis. Journal of Personality and Social

Sarason, I. G. (1988). Anxiety, self-preoccupation and

attention. Anxiety Research, 1,3
8.
Schwarzer, R. (Ed.) (1986). Self-related cognitions in

anxiety and motivation. Hillsdale,
NJ: Erlbaum.
Schwarzer, R., Jerusalem, M., & Stiksrud, A. (1984). The

developmental relationship
between test anxiety and helplessness. In H. M. van der

Ploeg, R. Schwarzer, & c. D.
Spielberger (Eds.), Advances in test anxiety research

(Vol. 3, pp. 73-79). Lisse, The
Netherlands/Hillsdale, NJ: Swets & Zeitlinger/Erlbaum.
Schwarzer, R., & Wicklund, R. A. (Eds.) (1991). Anxiety

and self-focused attention.
London: Harwood Academic Publishers.
Seligman, M. E. P. (1975). Helplessness. On depression.

development and death. San
Francisco: Freeman.
Smith, C. A., & Ellsworth, P. C. (1985). Patterns of

cognitive appraisal in emotion.
Smith, C. A., & Ellsworth, P. C. (1987). Patterns of

appraisal and emotion related to
taking an exam. Journal of Personality and Social

Spielberger, C. D. (1972). Conceptual and methodological

issues in anxiety research. In
C. D. Spielberger (Ed.), Anxiety: Current trends in theory

and research (Vol. 2, pp.
481-493). New York: Academic Press.
Spielberger, C. D. (1985). Anxiety, cognition and affect: A

state-trait perspective. In
H. Tuma & 1. Maser (Eds.), Anxiety and anxiety disorders

(pp. 171-182). Hillsdale,
NJ: Erlbaum.
Wine, 1. D. (1980). Cognitive-attentional theory of test

anxiety. In I. G. Sarason (Ed.),
Test anxiety: Theory. research, and applications (pp.

349-385). Hillsdale, NJ:
Erlbaum.
Wine, 1. D. (1982). Evaluation anxiety: A

cognitive-attentional construct. In H. W.
Krohne & L. Laux (Eds.), Achievement. stress. and anxiety

(pp. 207-219).
Wood, R. E., & Bandura, A. (1989). Impact of conceptions

of ability on self-regulatory
mechanisms and complex decision making. Journal of

Wortman, C. B., & Brehm, 1. W. (1975). Responses to

uncontrollable outcomes: An
integration of reactance theory and the learned

helplessness model. In L. Berkowitz
(Ed.), Advances in experimental social psychology (Vol. 8,

pp. 277-336). New York:
Academic Press. This page intentionally left blank

IV Self-Efficacy and Health Behaviors
Ajzen, I. (1988). Attitudes. personality. and behavior.

Milton Keynes: Open University
Press.
Ajzen, I., & Madden, J. T. (1986). Prediction of


Experimental Social
Annis, H. M., & Davis, C. S. (1988). Assessment of

expectancies. In D. M. Donovan &
G. A. Marlatt (Eds.), Assessment of addictive behaviors

(pp. 84-111). New York:
Guilford Press.
Baer, J. S., & Lichtenstein, E. (1988). Cognitive

assessment. In D. M. Donovan, &
Marlatt, G. A. (Eds.), Assessment of addictive behaviors

(pp. 104-110). New York:
Guilford Press.


Prentice-Hall.

Anxiety Research, 1, 77-98.
Bandura, A. (1989). Human agency in social cognitive

theory. American Psychologist,
44(9), 1175-1184.
Bandura, A. (1990). Perceived self-efficacy in the exercise

of control over AIDS infec
tion. Medienpsychologie, 2, 23-43.
Bandura, A. (1991). Self-efficacy mechanism in

physiological activation and health
promoting behavior. In J. Madden (Ed.), Neurobiology of

learning, emotion and
affect (pp. 229-270). New York: Raven Press.
Bandura, A., Cioffi, D., Taylor, C. 8., & Brouillard, M.

E. (1988). Perceived self
efficacy in coping with cognitive stressors and opioid

activation. Journal of

Micro-analysis of action and fear

Bandura, A., Taylor, C. 8., Williams, S. L., Mefford, I.

N., & Barchas, 1 D. (1985).

Barling, 1, & Abel, M. (1983). Self-efficacy and tennis

performance. Cognitive

threat seriousness and personal
efficacy upon intentions and behavior. Journal of Applied

Social Psychology, 11,
401-415.
Becker, H. M. (1974). The health belief model and personal

health behavior. Thorofare,
NJ: Slack.
Becker, M. H., & Rosenstock, 1. M. (1987). Comparing

social leaming theory and the
health belief model. In W. 8. Ward (Ed.), Advances in

health education and promo
tion (Vol. 2, pp. 245-249). Greenwich, CT: JAI Press.
Bernier, M., & Avard, J. (1986). Self-efficacy, outcome

and attrition in a weight reduc
tion program. Cognitive Therapy and Research, 10,319-338.
Chambliss, C. A., & Murray, E. 1 (1979). Efficacy

attribution, locus of control, and
weight loss. Cognitive Therapy and Research, 3,349-353.
Cohen, S., Lichtenstein, E., Mermelstein, R., Kingsolver.

K., Baer, 1. S .• & Kamarck,
T. W. (1988). Social support interventions for smoking

cessation. In B. H. Gottlieb
(Ed.), Marshaling social support. Formats, processes, and

effects (pp. 211-240).
Beverly Hills, CA: Sage.
Colletti, G., Supnick, 1 A., & Payne, T. 1 (1985). The

smoking self-efficacy ques
tionnaire (SSEQ): Preliminary scale development and

validation. Behavioral
Assessment, 7, 249-260.
Condiotte, M. M., & Lichtenstein, E. (1981). Self-efficacy

and relapse in smoking ces
sation programs. Journal of Consulting and Clinical

Critchlow, B. (1986). The powers of John Barleycorn:

Beliefs about the effects of alco
hol on social behavior. American Psychologist, 41, 751-764.
Critchlow Leigh, 8. (1989). In search of the seven

dwarves: Issues of measurement and
meaning in alcohol expectancy. Psychological Bulletin,

105,361-373.
De Vries, H., Dijkstra, M., & Kok, G. (1989, July).

Self-efficacy as a determinant of the
onset of smoking and interventions to prevent smoking in

adolescents. Paper present
ed at the First European Congress of Psychology. Amsterdam,

The Netherlands.

Self-efficacy: The third factor
besides attitude and subjective norm as a predictor of

De Vries, H., & Kok, G. J. (1986). From determinants of

smoking behavior to the impli
cations for a prevention programme. Health Education

Research, I, 85-94.
Desharnais, R., Bouillon, J., & Godin, G. (1986).

Self-efficacy and outcome expecta
tions as detenninants of exercise adherence. Psychological

Reports, 59, 1155-1159.
DiClemente, C. c., Prochaska, J. 0., & Gibertini, M.

(1985). Self-efficacy and the
stages of self-change of smoking. Cognitive Therapy and

Research, 9, 181-200.
Donovan, D. M., & Marlatt, G. A. (Eds.). (1988).

Assessment of addictive behaviors.
New York: Guilford Press.
Dzewaltowski, D. A. (1989). Toward a model of exercise

motivation. Journal of Sport
& Exercise Psychology, 11,251-269.
Eiser, 1. R. (1983). Smoking, addiction and

decision-making. International Review of
Applied Psychology, 32 , 11-28.
Eiser, 1. R., & Sutton, S. R. (1977). Smoking as a

subjectively rational choice. Addictive
Behaviors, 2, 129-134.
Ewart, C. K., Gillian, R. E., Kalemen, M. H., Manley, S.

A., & Kalemen, M. D. (1986).
Usefulness of self-efficacy in predicting overexertion

during programmed exercise in
coronary artery disease. American Journal of Cardiology,

57,557-561.
Fishbein, M. (1982). Social psychological analysis of

smoking behavior. In J. R. Eiser
(Ed.), Social psychology and behavioral medicine (pp.

179-197). Chichester: Wiley
& Sons.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude,

intention, and behavior: An introduc
tion to theory and research. Reading, MA: Addison-Wesley.
Garcia, M. E., Schmitz, 1. M., & Doerfler, L. A. (1990). A

fine-grained analysis of the
role of self-efficacy in self-initiated attempts to quit

smoking. Journal of Consulting
and Clinical Psychology, 58,317-322.
Gilchrist, L. D., & Schinke, S. P. (1983). Coping with

contraception: Cognitive and
behavioral methods with adolescents. Cognitive Therapy and

Research, 7, 379-388.
Glynn, S. M., & Rudennan, A. J. (1986). The development

and validation of an eating
self-efficacy scale. Cognitive Therapy and Research, 10,

403-420.
Godding, P. R., & Glasgow, R. E. (1985). Self-efficacy and

outcome expectations as

Heckhausen, H. (1991). Motivation and action. Berlin:

Springer.
Heckhausen, H., & Kuhl, 1. (1985). Fonn wishes to actions:

The dead ends and short
cuts on the long way to action. In M. Frese & J. Sabini

(Eds.), Goal-directed
behavior: The concept of action in psychology (pp.

134-159). Hillsdale, NJ: Erlbaum.
Janz, N. K., & Becker, M. H. (1984). The health belief

model: A decade later. Health
Education Quarterly, 11, 1-47.
Kaplan, R. M., Atkins, C. 1., & Reinsch, S. (1984).

Specific efficacy expectations
mediate exercise compliance in patients with COPD. Health

Kok, G., De Vries, H., Mudde, A. N, & Strecher, V. J.

(1990). Planned health education
and the role of self-efficacy: Dutch research. Health

Education Research, 5.
Kuhl, J. (1983). Motivation, Konflikt und

Handlungskontrolle [Motivation, conflict, and
action control]. Berlin: Springer.
Kuhl, J. (1985). Volitional mediators of

cognition-behavior consistency: Self-regulatory
processes and action versus state orientation. In 1. Kuhl

& 1. Beckmann (Eds.), Action
control: From cognition to behavior (pp. 101-128). New
York: Springer.
Kuhl, 1. (1986). Motivation and infonnation processing. A

new look at decision mak
ing, dynamic change, and action control. In R. M.

Sorrentino & E. T. Higgins (Eds.),
Handbook of motivation and cognition (pp. 404-434). New

York: Wiley.
Kuhl, 1., & Krasna, K. (1990). Self-regulation and

metamotivation: Computational
mechanisms, development, and assessment. In R. Kanfer, P.

L. Ackerman, &
R. Cudeck (Eds.), Learning and individual differences (pp.

343-374). Hillsdale, NJ:
Erlbaum.
Lazarus, R. S. (1991). Emotion and adaptation. London:


Springer.
Lazarus, R. S., & Folkman, S. (1987). Transactional theory

and research on emotions
and coping. European Journal of Personality, I, 141-170.
Levinson, R. A. (1982). Teenage women and contraceptive

behavior: Focus on self
efficacy in sexual and contraceptive situations.

Unpublished Doctoral Dissertation.
Litt, M. D. (1988). Self-efficacy and perceived control:

Cognitive mediators of pain tol
erance. Journal of Personality and Social Psychology, 54,

149-160.
Maddux, J. E., Norton, L. W., & Stoltenberg, C. D. (1986).
Self-efficacy expectancy,
outcome expectancy, and outcome value: Relative effects on

behavioral intentions.
Maddux, 1. E., & Rogers, R. W. (1983). Protection

motivation and self-efficacy: A
revised theory of fear appeals and attitude change.

Journal of Experimental Social
Manning, M. M., & Wright, T. L. (1983). Self-efficacy

expextancies, outcome expect
ancies, and the persistence of pain control in childbirth.

Marlatt, G. A. (1985). Cognitive factors in the relapse

process. In G. A. Marlatt & J. R.
Gordon (Eds.), Relapse prevention: Maintenance strategies

in the treatment of
addictive disorders (pp. 128-200). New York: The Guilford

Press.
Marlatt, G. A., Baer, 1. S., Donovan, D. M., & Kivlahan,

D. R. (1988). Addictive behav
iors: Etiology and treatment. Annual Review of Psychology,

39, 223-252.
Marlatt, G. A., & Gordon, J. R. (Eds.). (1985). Relapse

prevention: Maintenance strate
gies in the treatment of addictive behaviors. New York:

Guilford.
McKusick, L., Coates, T. J., Morin, S. F., Pollack, L., &

Hoff, C. (1990). Longitudinal
predictors of reductions in unprotected anal intercourse

among gay men in San
Francisco: The AIDS behavioral research project. American

Journal of Public
Health, 80, 978-983.
Meyerowitz, B. E., & Chaiken, S. (1987). The effect of

message framing on breast self
examination attitudes, intentions. and behavior. Journal

Mischel, H. N., & Mischel, W. (1983). The development of

children's knowledge of
self-control strategies. Child Development, 54,603-619.
Mischel, W. (1966). Theory and research on the antecedents

of self-imposed delay of
reward. In B. A. Maher (Ed.), Progress in experimental

personality research (Vol. 3,
pp. 85-132). New York: Academic Press.
Mischel, W. (1973). Toward a cognitive social learning

reconceptualization of personal
ity. Psychological Review, 80, 252-283.
Mudde, A., Kok, G., & Strecher, V. (1989, July).

Self-efficacy and success expectancy
as predictors of the cessation of smoking. Paper presented

at the First European
Congress of Psychology. Amsterdam, The Netherlands.
Mullen, P. D., Hersey, J. C., & Iverson, D. C. (1987).

Health behavior models com
pared. Social Science and Medicine, 24, 973-983.
O'Leary, A. (1985). Self-efficacy and health. Behavioral

Research Therapy, 23, 437
451.
O'Leary, A., Shoor, S., Lorig, K., & Holman, H. R. (1988).
A cognitive-behavioral
treatment for rheumatoid arthritis. Health Psychology, 7,

527-542.
Paulussen, T., Kok, G. 1., Knibbe, R., & Kramer, T. (1989).

AIDS en intraveneus drug
gebruik (AIDS and IV-drug use). Tijdschrift voor Sociale

Gezondheitszorg [Dutch
Journal of Social Health Care], 68,129-136.
Rippetoe, P.A., & Rogers, R. W. (1987). Effects on

components of protection motiva
tion theory on adaptive and maladaptive coping with a

health threat. Journal of
Personality & Social Psychology, 52,596-604.
Rogers, R.W. (1975). A protection motivation theory of fear

appeals and attitude
change. Journal of Psychology, 91, 93-114.
Rogers, R. W. (1983). Cognitive and physiological processes

in fear appeals and atti
tude change: A revised theory of protection motivation. In

J. R. Cacioppo & R. E.
Petty (Eds.), Social psychology: A sourcebook (pp.

153-176). New York: Guilford.
Rogers, R. W. (1985). Attitude change and information

integration in fear appeals.
Psychological Reports, 56, 179-182.
Rosenstock, I. M. (1966). Why people use health services.

Milbank Memorial Fund
Quarterly, 44, 94.
Rosenstock, I. M., Strecher, V. 1., & Becker, M. H.

(1988). Social learning theory and
the Health Belief Model. Health Education Quarterly, 15,
175-183.
Sallis, 1. F., Haskell, W. L., Fortmann, S. P., Vranizan,

K. M., Taylor, C. B., &
Solomon, D. S. (1986). Predictors of adoption and

maintenance of physical activity in
a community sample. Preventive Medicine, 15,331-341.
Scheier, M. F., & Carver, C. S. (1985). Optimism, coping,


Scheier, M. F., & Carver, C. S. (1987). Dispositional

optimism and physical well-being:
The influence of generalized outcome expectancies on

health. Journal of Personality,
55,169-210.
Scheier, M. F., Weintraub, 1. K., & Carver, C. S. (1986).

Coping with stress: Divergent
strategies of optimists and pessimists. Journal of

51, 1257-1264.
Schifter, D. B., & Ajzen, I. (1985). Intention, perceived

control, and weight loss: An
application of the theory of planned behavior. Journal of

Schwarzer, R. (1981). Strefi. Angst und Hilflosigkeit

(Stress, anxiety, and helplessness].
Stuttgart: Kohlhammer.
Schwarzer, R. (1992). Psychologie des Gesundheitsverhaltens

[Psychology of health
behavior). Gottingen: Hogrefe.

Seydel, E., Taal, E., & Wiegman, O. (1990). Risk-appraisal,
outcome and self-efficacy
expectancies: Cognitive factors in preventive behavior

related to cancer. Psychology
and Health, 4,99-109.
Slater, M. D. (1989). Social influences and cognitive

control as predictors of self
efficacy and eating behavior. Cognitive Therapy and

Research, 13,231-245.
Strecher, V. 1., DeVellis, B. M., Becker, M. H., &

Rosenstock, I. M. (1986). The role of
self-efficacy in achieving health behavior change. Health

Education Quarterly, /3,
73-91.
Sutton, S. (1989). Smoking attitudes and behavior:

Applications of Fishbein and
Ajzen's Theory of Reasoned Action to predicting and

understanding smoking
decisions. In T. Ney & A. Gale (Eds.), Smoking and human

behavior (pp. 289-312).
Chichester: Wiley.
Taylor, C. B., Bandura, A., Ewart, C. K., Miller, N. H., &

DeBusk, R. F. (1985). Exer
cise testing to enhance wives' confidence in their

husbands' cardiac capability soon
after clinically uncomplicated acute myocardial infarction.

American Journal of
Cardiology, 55,635-638.
Velicer, W. F., DiClemente, C. C., Rossi, J. S., &

Prochaska, 1. O. (1990). Relapse situ
ations and self-efficacy: An integrative model. Addictive

Behaviors, 15, 271-283.
Expectations and performance: An

of Sport Psychology, 1,320
331.
Weinberg, R. S., Gould, D., Yukelson, D., & Jackson, A.

(1981). The effect of pre
existing and manipulated self-efficacy on competitive

muscular endurance task.
Journal of Sport Psychology, 4,345-354.
Weinberg, R. S., Hughes, H. H., Critelli, J. W., England,

R., & Jackson, A. (1984).
Effects of preexisting and manipulated self-efficacy on

weight loss in a self-control
program. Journal of Research in Personality, 18,352-358.
Weinberg, R. S., Yukelson, D., & Jackson, A. (1980).

Effects of public and private
efficacy expectations on competitive performance. Journal

of Sport Psychology, 2,
340-349.
Weiner, B. (1980). The role of affect in rational

(attributional) approaches to human
motivation. Educational Researcher, 9, 4-11.
Weinstein, N. D. (1988). The precaution adoption process.

Health Psychology, 7, 355
386.
Weinstein, N. D. (1989). Effects of personal experience on

self-protective behavior.
Psychological Bulletin, 105,31-50.

Levine, S .• & Raska, K.
stressors on immune func
tion. Journal of Personality and Social Psychology, 59,

1082-1094.
Wills, T. A. (1990). Multiple networks and substance use.

Journal of Social and
Wilson, D. K., Wallston, K. A., & King, J. E. (1990).

Effects of contract framing, moti
vation to quit, and self-efficacy on smoking reduction.

Journal of Applied Social
Wood, 1. (1989). Theory and research concerning social

comparisons of personal attri
butes. Psychological Bulletin, 106,231-248.
Wurtele, S. K., & Maddux, 1. E. (1987). Relative

contributions of protection motivation
theory components in predicting exercise intentions and

behavior. Health
Yalow, E. S., & Collins 1. L. (1987). Self-efficacy in

health behaviour change: Issues in
measurement and research design. In W. B. Ward (Ed.),

Advances in health educa
tion and promotion (Vol. 2, pp. 181-199). Greenwich, CT:

JAI Press. Author Notes The author is grateful to Albert
Bandura, Charles Carver, and two anony
mous referees for their valuable comments on an earlier

draft of this chapter. This page intentionally left blank
SELF -EFFICACY AND A TTRIBUTION THEORY IN HEALTH
EDUCATION Gerjo Kok, Dirk-Jan Den Boer, Hein De Vries,
Frans Gerards, Harm J. Hospers, and Aart N. Mudde
Self-efficacy is defined as the estimation of the person
about his/her ability to perform a specific behavior. The
important role of selfefficacy in health education is
argued, illustrated with Dutch research, which shows that
self-efficacy is an important determinant of health
behavior, of future health behavior, and of health
behavior change. It will be discussed how attribution
theory is related to self-efficacy theory and how insights
from attribution theory can be applied in health behavior
interventions. Programs to improve self-efficacy by
attributional retraining have been developed successfully,
mostly based on relapse prevention theory.
Health education is a fonn of planned behavior change (see

Green & Lewis,1986;
Kok, 1988). Research shows that the effectiveness of health

education activities
is detennined by the quality of this planning process

(Jonkers, De Haes, Kok,
Liedekerken, & Saan, 1988): analysing the problem,

analysing the related behav
iors, analysing the detenninants of that behavior,

developing interventions to
change that behavior, organising implementation, and

finally evaluating of the
effects on the problem or, at least, the behavior. An

important step in this plan
ning process is the analysis of the determinants of the

target behavior. We will
focus on self-efficacy as a detenninant of behavior and as

a predictor of future
behavior and behavioral change, using illustrations from

our own research. We
will then suggest strategies to improve self-efficacy,

based on attribution theory
and relapse prevention theory. DETERMINANTS OF BEHAVIOR

Social psychological theory and research, fundamental as
welJ as applied, has
traditionally provided an important contribution to the

understanding of the deter
minants of behavior. Fishbein and Ajzen (1975) integrated a
series of models
from attitude theory and social influence theory in their

model of reasoned
behavior. They stimulated a whole field of fundamental and

especially applied
research (Ajzen & Fishbein,1980; De Vries & Kok, 1986).

Gradually it became
clear that the Fishbein and Ajzen model was very useful,
but limited. The model
could not sufficiently account for the important role of

behavioral costs and bar
riers in performing the behavior. In theory, costs and

barriers should appear as
beliefs about performing the behavior, but in practice

researchers found better
prediction of behavior when costs and barriers were

measured independently of
the attitude. In addition to the model of reasoned

behavior, researchers focussed
on Bandura's concept of self-efficacy (Bandura, 1986;

Strecher, DeVel1is,
Becker, & Rosenstock, 1986) as a determinant of behavior.

Self-efficacy is the
estimation of the person about his/her ability to perform

a specific behavior in a
specific situation. Self-efficacy expectations are based

on own experience with
the behavior (and especially the attributions that people

make about success and
failure), observations of others, persuasion by others, and

physiological informa
tion (e.g., nervousness). Self-efficacy has to be

distinguished from outcome
expectancy, the latter being an estimation of the

effectiveness of the behavior to
reach a desired goal. For a smoker, outcome expectancy is

the estimation of the
improvement in health and life by stopping smoking, while

self-efficacy is the
estimation of own ability to really give up smoking.

Several researchers reported
an improvement in the prediction of behavior by combining

attitudes, social
norms, and self-efficacy (Ajzen, 1987; De Vries, Dijkstra

& Kuhlman, 1988).
This "Determinants of Behavior Model" is represented in

Figure I.
Figure 1 Detenninants of behavior model. Attitudes,

social norms and self-efficacy predict the intention, which
in turn
predicts the behavior. External variables (outside the

model), like demographic
ones, are supposed to influence behavior via the three

determinants and the inten
tion. Between intention and behavior there can be barriers

or lack of skills. Atti
tudes, social norms and self-efficacy can be measured in

advance, while barriers
and skills play a role when the behavior is actually

performed. The intention
predicts the behavior, but the model also indicates that

self-efficacy is an estima
tion of the skills that are needed in the actual situation

and the possibilities to
overcome barriers. For that reason an influence from

self-efficacy via skills on
behavior is proposed. The actual performance of the
behavior leads to a feedAttitude Barriers Intention
Behavior Skills
External
variables Social norms Self-efficacy
back process that influences in tum the three determinants.

Attitudes, social
norms and self-efficacy are not completely independent from

each other. Mostly
the correlations between the three are substantial. All

three determinants can be
characterised as beliefs. But empirical data indicate that

a separate measurement
of each determinant improves the prediction of behavior

significantly. There is as yet no tradition in the
measurement of self-efficacy. Bandura
(1986) has argued that self-efficacy expectations vary

along dimensions of mag
nitude, generality, and strength. This implies that

self-efficacy estimates must be
viewed as situation dependent. The perception of subjects'

ability to perform a
certain behavior is determined, and varies, by the

perceived task difficulty and
the situation. Measurement of perceived difficulty does not

necessarily imply an
estimation of ability. A person can see a behavior in a

certain situation as very
difficult, but also as something that can be done.

Measuring ability in relation to
situational aspects, however, is likely to include

perceived difficulty levels of
both situation and behavior. A person asked to rate his

confidence of being able
to perform a certain behavior in a certain situation, will

probably weigh ability
against task difficulty and situation difficulty. This

leads to the conclusion that
instruments to measure perceived ability ("Do you think

you are able to ... ")
assess both the dimensions strength and magnitude.

Including a sample of rele
vant situations provides an assessment of the dimension

generality. In the following we will present some of our
studies about the role of self
efficacy as a determinant of behavior and behavioral

change (for a review of the
international literature, see Strecher et al.,1986). Most

of these studies are cross
sectional, measuring the determinants and the behavior at

the same time. [n addi
tion we will present two of our studies that are

longitudinal, predicting future
behavior from former measures of determinants. This makes

it possible to infer
causal relationships. Dutch Research on Self-Efficacy as a

Detenninant of Behavior De Vries (1989) analyzed the
determinants of the onset of smoking in youth.
Eighty-five third grade Dutch pupils of various secondary

schools participated in
his study, 40% male and 60% female, with age varying
between 14 and 17 years.
Sixteen questions assessed the attitude by focussing on

both shortand long-term
consequences, and personal, social and health consequences.

Social norms were
assessed by nine questions about the norms of parents,

siblings, peers and adults.
The nine questions on self-efficacy comprised: finding it

difficult/easy not to
smoke when friends smoke, explaining to other people that

"} do not want to
smoke," being able to refuse a cigarette when offered, when

offered by parents,
when offered by friends, in spite of being called a

coward, being able to stop
smoking when wanting to, knowing a reason to refuse a

cigarette, succeeding in
becoming/staying a non-smoker. One item measured the

intention and one item
measured the behavior. The nine self-efficacy questions

formed a reliable scale,
with Cronbach's alpha .80. Figure 2 shows the correlations

and multiple correla
tions of the determinants with intention and behavior.

Attitude .66 1.47 R-.79 .74 R-.BOI Social norms
Intention Behavior .71 I Sel f-efficacy I .66
Figure 2 Correlations and mUltiple correlations of the

three determinants with intention and smoking behavior (N
= 85, all correlations are significant p < .01). A
hierarchical regression analysis showed that self-efficacy
had a unique
contribution in the prediction of the intention, when

added after attitude and
social norm. Self-efficacy explained another 15% of the

variance in the inten
tion. Moreover, self-efficacy had a unique contribution in

the prediction of the
behavior, when added after the intention, explaining

another 9%. So in line with
our expectations, De Vries found an influence from

self-efficacy on the intention
independent of attitude and social norm, and an influence
on behavior independ
ent of the intention. The latter is probably the result of

the relation between self
efficacy estimations and the actual skills in performing

the behavior and over
coming barriers. On seven of the nine self-efficacy items

there was a significant
difference between the smokers and the non-smokers (see De

Vries, 1989). We have presented a study that shows a
relation between attitudes, social
norms and self-efficacy on the one hand, and intention and

behavior on the other.
A number of other studies have found the same result.

These were cross-sec
tional studies that cannot show any causal relationship.

We know that attitudes
cause behavior, but behavior in tum causes attitudes. The

same applies to social
norms. With respect to self-efficacy, the first assumption

is that self-efficacy is
the result of behavior. In our model, however, we have

assumed that self
efficacy can also cause behavior. In the following we will

present two studies
that are longitudinal and that show that self-efficacy is a

determinant of future
behavior. The reported study by De Vries (1989) was part

of a research program about
the prevention of smoking in youth, with a series of

measures over time. The
attitude, social norm and self-efficacy scores of the

control group at the first
measurement (Time I), have been used to predict intention
and behavior at a
following measurement, one year later (Time 2). The

correlations of the determ
inants at Time I with intention and behavior at Time 2 are

depicted in Figure 3.
Figure 3 Correlations and multiple correlations of the

three determinants at Time 1 with (I) the intention at
Time 2, and (2) with smoking behavior at Time 2 (N = 600;
all correlations are significant P $: .0 I). The results
of a regression analysis showed that self-efficacy at Time
1 was
the best predictor of smoking intention at Time 2 one year

later, explaining 24%
of the variance. Self-efficacy at Time 1 also had a unique

contribution in the
prediction of behavior at Time 2, when added after the

intention Time 1. This
study showed that self-efficacy, measured at a certain

time, can predict future
intentions and behavior. The study of Mudde, Kok, and

Strecher (1989) focussed on self-efficacy as a
predictor of success in the cessation of smoking. Subjects

were 123 participants
in a three-week "Stop Smoking" program. The most important

principles of the
program were: (a) quitting at the first meeting, (b)

concentrating on one's own
potential to fight addiction through willpower, and (c)

preparation for the physi
ological consequences of quitting. Enhancing

self-efficacy was not the focus of this program. Although
certain
components of the program are likely to enhance

self-efficacy, no specific atten
tion was paid to skills training or obtaining adequate
coping responses. The pos
sibility of relapse and how to handle it when it occurs

received minimal attention.
The program consisted of seven meetings, spread over three

weeks. Both before
and after treatment and after follow-up periods of six

weeks and one year, self
efficacy and smoking behavior were measured by

questionnaire. Two measures
related to self-efficacy were used in this study: 1. A

one-item perceived ability measure. Subjects were asked to
rate their
perceived ability in resisting the urge to smoke in every

possible situation. 2. The "Smoking Self-Efficacy
Questionnaire" (SSEQ), developed by
Coletti, Supnick and Payne (1985). Subjects were asked to

rate their perceived
ability in resisting the urge to smoke in seventeen

different situations. .44 Attitude T1 .37 R-.54 Social
norm T1 Intention T2 .49 Self-efficacy T1 .50 Intention T1
.64 R-.65 Behavior T2 At the pre-treatment self-efficacy
measure, respondents were told to imagine
that they were quitting without professional assistance, to

minimize the effect of
program-efficacy expectations. Assessment of smoking

behavior was realised by
self-report. The cutting point between smokers and quitters

was set at one ciga
rette or more during the last seven days before

measurement. Once a participant
was missing at a follow-up, that person was treated as a

smoker in the following
analyses. Success rates of the program were 54% after

treatment, 44% after six-weeks
follow-up, and 27% after one-year follow-up. Schwartz
(1987) reports median
success rates for group interventions of 24% to 36% after

one-year follow-up, so
the result of this treatment is comparable. There are no

differences between the
groups of quitters and smokers after one year, with

respect to any measure on the
pretest. We will focus on the role of self-efficacy after

treatment in predicting
smoking behavior after one year. Mudde et a1. (1989)

divided the participants
that were successful after the treatment in three

success/failure groups: A: post-treatment success,
post-six-weeks success, post-one-year success: B:
post-treatment success, post-six-weeks success,
post-one-year failure; C: post-treatment success,
post-six-weeks failure, post-one-year failure. Mudde et
al. predicted the membership of Groups A, Band C after one
year,
from the self-efficacy scores at the post-treatment

measure, and the increase in
self-efficacy during treatment (see Table 1)

(post-treatment failures were left out
because their post-treatment self-efficacy scores will

artificially be lower than
those of the other groups). The post-treatment level of

self-efficacy and the in
crease of self-efficacy during treatment, as measured by

the one-item perceived
ability measure, were predictors for success and failure

after one year. In the
SSEQ measures there is a slight trend. Again we find that

self-efficacy at a cer
tain time can predict behavior change in the future, in

this case the long-term
effects of a smoking cessation program.
Table I
Self-Efficacy Scores at Post-Treatment Measures and Success

and Failure Groups After
One Year (N = 66) Post-Treatment Score Increase During

Treatment One-Item SSEQ One-Item SSEQ
Group A 3.8 4.5 2.2 2.2
Group B 3.4 4.1 1.6 1.9
Groupe 3.2 4.4 1.2 1.7
p < .05 .14 .01 .17 Improving Self-Efficacy by Health

Education We have presented research that shows the
contribution of self-efficacy in
the determination of behavior. We have also presented

research showing that
self-efficacy is predictive for future success and failure

in behavioral change.
However, the program in the study by Mudde et al. (1989)

was not specifically
meant to improve self-efficacy. An important question for

health education is:
Can we improve self-efficacy and thus stimulate the desired

behavior change?
This question cannot fully be answered positively at this

moment, but we do have
promising evidence. During the last decade there has been

an increase in tech
niques based on attribution theory that can be used to

induce people to change
their behavior and based on attributional retraining and

self-efficacy improve
ment (Weary, Stanley, & Harvey, 1989). We will now discuss

attribution theory
and show how this theory is related to self-efficacy
theory. we will also show
how attributional insights can be applied in health

behavior change programs. ATTRmUTION THEORY Figure 4 is a
schematic depicting Weiner's attributional model (1985,
1986).
The model indicates that when outcomes are negative,

unexpected, or important a
causal search is started, resulting in causal ascriptions

that try to explain the out
comes. For example, a person who attempts to quit smoking

but fails might attri
bute this outcome to a variety of reasons, e.g., low

effort, the difficulty of the
task, or nicotine dependency. According to attribution

theory, perceived causal
reasons have an underlying dimensional structure (Heider,

1958; Kelley, 1967). Theoretically, three attributional
dimensions have been distinguished. The
first dimension, locus of causality, reflects the extent

to which previous outcomes
are attributed to causes either internal or external to

the person. The second
dimension, stability, reflects the extent to which

previous outcomes are attributed
to stable or unstable causes. Stable causes refer to

perceptions that a failure or
success was due to immutable, unalterable causes. Unstable

causes refer to per
ceptions that a failure or a success was due to causes

that were mutable. The
third dimension, controllability, reflects the extent to

which previous outcomes
are attributed to controllable or uncontrollable causes.

Empirical support is
strongest for the dimensions locus of causality and

stability, the support for the
dimension controllability is weakest (Hews tone & Antaki,

1988; Weiner, 1986).
Attribution theory contends that causal ascriptions

influence cognitions and emo
tions related to success and failure. We will first

describe the cognitive
component and then turn to the affective component of

attribution theory. Cognitive Component of Attribution
Theory Weiner's model suggests that expectancy of success
is determined by the
perceived stability of the causes for success or failure. A

person attributing a
success to a stable cause (e.g., ability) will have a

higher expectancy of success
when having to perfonn the same task again, compared to

somebody who attri
butes a success on the same task to an unstable cause

(e.g., luck). After failure
this effect is reversed. The rationale behind this

assumption is that when there is
no reason to expect the cause of failure to change, the

second time the task will
be perfonned the cause responsible for failure will still

be present. If the cause
for failure was unstable, there is no reason to expect the

cause to be present the
next time one performs the task.
Figure 4 Schematic overview of Weiner's attribution theory.

Furthennore, it is assumed that a lowered expectancy of
success leads to less
adaptive task behavior. Because of the lowered expectation

of success, persons
will invest less energy in the task at hand, because they

perceive a lower likeli
hood of succeeding. It is this lack of energy invested in

the task which causes the
low perfonnance. This line of reasoning very closely

parallels Bandura's ideas
about self-efficacy. He assumes that a low self-efficacy

leads to avoidance of the
task at hand. Self-efficacy estimates resemble expectancies

of success in the
respect that both concepts are related to the estimate

people make about the like
lihood that a certain outcome or goal will be attained.

Furthennore, both
estimates are based on a cognitive appraisal of past

experiences (Bandura, 1986,
p. 349; Weiner, 1986, p. 181). Several studies on

attributional processes have
used self-efficacy ratings as a measure of expectancy of

success, and some
authors claim that self-efficacy and expectancy of success

are identical concepts
(e.g., Kirsch, 1985, 1986). outcome emotion information

attr ibution stabil i ty locus controllabil ity expectation
of success emotions behavioral consequences outcome Support
for the cognitive component of this theory can be found,
among
others, in a study by Eiser and Van der Pligt (1986) on

smoking behavior, and a
study by Hospers, Kok, and Strecher (1990) on weight

reduction. Hospers et aI.
tested the cognitive component of the attributional model

on 158 subjects who
participated in a weight reduction program. They measured
the number of previ
ous attempts to lose weight, stability of attributions for

previous failure, expect
ancy of success and goal attainment. Results of this study

are shown in Figure 5.
Figure 5 Path analysis with goal attainment as outcome

variable; bold arrows indicate significant ~-weights (p <
.01). As hypothesized, goal attainment is positively
associated with success
expectancy which, in tum, is negatively associated with

stability. Furthermore,
stability was positively related to the number of former

trials. Also as hypothe
sized, there was no significant association between the

number of previous
attempts and goal attainment, or stability and goal

attainment. These results sug
gest that it is not the number of failures people

experience that is important, but
the way people interpret these failures. The more stable

the causes attributed to
failure, the lower the expectancy of success for the next

attempt, and hence lower
goal attainment. As the results of the above-mentioned

studies are all correlational, one should
be careful in interpreting these results. These studies do

not show that because
subjects attribute their previous failures at losing weight

(or quitting smoking
[Eiser & Van der Pligt, 1986]) to stable causes they do

not achieve their goals.
Correlational studies can only show that a relationship

between variables exists,
they can never show that a change in one variable is the
cause of the change in
another variable. Only some studies have shown a causal

relationship between
attributional change and behavioral change (Den Boer,

Meertens, Kok, & Van
Knippenberg, 1990). Affective Component of Attribution

Theory The relation between stability of attributions,
expectation of success and
behavior can be referred to as the cognitive component of

Weiner's theory. The
theory also consists of an affective component: the

relation between perceptions .00 .22 numbe r of failures
stabi l i ty -.33 expectancy of success .37 goal attainment
-.02 -.12
of locus and controllability on the one hand and emotions

and behavior on the
other. It is assumed that different attributions lead to

different emotions. An
attribution of effort for failure (internal/controllable)

will lead to guilt, while an
attribution of ability (internal/uncontrollable) for

failure will lead to shame. [t is
assumed that there are two kinds of emotions: motivating

and debilitating. Moti
vating emotions (guilt for instance) will lead to a better

task performance while
debilitating emotions (shame) will lead to a worse task

performance. Attribution theory states that after failure
guilt (as a result of an attribution to,
for instance, effort) will lead to better task performance

than shame (as a result of
an attribution to, for instance, ability). This statement

makes the results found by
Clifford (1986) interesting. She presented subjects with a

scenario in which
students failed an exam because of lack of ability, choice

of a wrong strategy or
lack of effort. She found that subjects predicted the

highest rate of success for
those students who failed because of the choice of a wrong

strategy, no differ
ences where found between ability or effort attribution

condition. She explains
these results by stating that an extremely high or an

extremely low level of guilt
(associated with effort or ability respectively) will lead

to a worse performance
than a moderate level of guilt (associated with an

attribution to strategy). It seems that the relationship
between emotions and task behavior is less
clear than it is assumed to be in attribution theory. At

this point it is safe to
assume that a relationship between attributions and outcome

expectancies exists
that is related to task behavior. Different kinds of

research paradigms have
shown this to be true (Den Boer et al., 1990; Hospers et

aI., \990). However,
there remains doubt about the affective component of this

theory. Especially the
question of which emotions could be termed debilitating and

which motivating,
remains to be answered. Applying Attributional Theories

Earlier we stated that attributional techniques can be, and
are already, applied
to many different situations and problem areas. Examples of

attributional expla
nations can be found in problem areas as diverse as

loneliness (Anderson,
Horowitz, & French, 1983), alcoholism (McHugh, Beckman, &

Frieze, 1979),
smoking (Eiser & Van der Pligt, 1986), losing weight

(Hospers et al., 1990), cop
ing with critical life events like accidents, rape and

illness (Janoff-Bulman,
1979). While the above-mentioned studies are all

descriptive, there have also
been attempts at changing people's behavior by changing

their attributions (see
Forsterling, 1988, for a review). There have been attempts

to improve reading
skills of children (Fowler & Peterson, 1981), improve

arithmetic skills of
children (Schunk, 1984), improve the score of subjects on

anagram tasks
(Andrews & Debus, 1978), lower drop-out rates at

high-school, and improve aca
demic success (Wilson & Linville, 1982). All these studies

are based on the fol
lowing assumption: If different attributions lead to

different behavioral con
sequences, it should be possible to change the behavior of

people by changing the
attributions they make. As most research within this

paradigm is focused on attri
butions after failure, it is assumed that after failure an

attribution to effort leads to
"better" task behavior than an attribution to, for

instance, ability. One could also
say that an internal, stable and uncontrollable attribution

for failure should be
substituted by an internal, unstable and controllable
attribution (Figures 6 and 7).
Because of the change from stable to instable there is no

need for the subjects to
expect a renewed failure. Because of the controllability of

the cause they can
actively try to change it: They can invest more energy in

the task, hoping that it
will improve their performance. This implies that, because

of the change in attri
bution, they will have a higher expectation of success,

which in turn will increase
their task performance. Helping people lose weight is a

good example of how this
theory can be applied. Research shows that the cause people

perceive for their
failure is predominantly internal, stable and

uncontrollable (Hospers et al., 1990).
People often think that they are incapable of losing weight

because they perceive
no relation between the amount of food they eat and their

weight. They think that
they will get fat anyway, even if they eat small amounts of
low-calorie food. This
is a typical case of dysfunctional attribution. Changing

this stable, internal and
uncontrollable attribution to unstable, internal and

controllable will result in a
higher possibility of a successful attempt at losing weight

the next time. Another
possibility is that they do perceive the relation between

amount of food eaten and
their weight, but do not think that they can control their
behavior because they
feel they do not have the willpower to change this
behavior. Again an example of
a dysfunctional attribution.
Figure 6 Undesirable attributional sequence. failure sad

information ability stable internal uncontrollable
expectation of failure sham e avoidance of task/ less
effort Investment fsilure Applications of attributional
insights are plentiful (Den Boer, Meertens, Kok,
& Van Knippenberg, 1989; Forsterling, 1988; Schunk, 1984;

Wilson & Linville,
1982). There remains, however, a significant amount of

doubt about the causes
of the positive effects found in these reattribution

techniques. The fact that there
has been hardly any research in which different

attributions are compared with
each other makes it difficult to state which attributions

are "the best" to attribute
failure or success to. The fact that providing subjects

with a simple attributional
questionnaire leads to increases in perfonnances, sheds

some doubt on the neces
sity of complex reattribution techniques as well. Despite

these problems we will
discuss an approach to maintenance of behavioral change

that incorporates many
of the above-mentioned insights: relapse prevention theory.

RELAPSE PREVENTION
A theory that uses both the concepts of attribution and

self-efficacy is relapse
prevention theory. Relapse prevention theory explains why

people who try to quit
a certain addictive behavior (e.g., smoking or drinking)

often fail. Marlatt and
Gordon (1985) have elaborated a relapse prevention theory
that can be explained
in attributional and self-efficacy concepts. An overview of

the theory is presented
in Figure 8.
Figure 7 Desired attributiona! sequence. failure sad

information effort unstable internal controllable no
expectation of failure guilt/ anger higher persistence
success
Figure 8 The relapse-prevention model (adapted from Marlatt

& Gordon, 1985, p. 106). An important concept in this
theory is the so-called "high-risk situation." A
high-risk situation is a situation in which a person is

tempted to return to his or
her old habit. For a smoker who tries to quit smoking, a

meeting with friends
who smoke or a day of hard work might be a high-risk

situation. In order to cope
with such a situation the person needs a coping response.

This means that the
person needs to anticipate the situation and know what to

do when that situation
arises. According to the theory the absence of a coping

response will lead to de
creased self-efficacy and initial use of the substance,

resulting in a relapse. If
there is a relapse, it is important to consider what

caused this lapse, and what can
be done to avoid such lapses in the future. The use of the

word "caused" in the
previous sentence already suggests the importance of

attribution theory. There
are causes that a person should not attribute to (for

instance stable causes like
ability or willpower) in this situation because that would
result in an even lower
self-efficacy. Attributing to stable, internal causes will

also lead to shame and
perceived loss of control. The smoker who has decided to

quit and finds himself high risk situation
presence of a
coping response absence of a coping response positive

outcome expectancies
increased
self efficacy decreased self efficacy
decreased
probability of
relapse us e of substance resolution not to use substance

sense of failure self attribution guilt/ shame perceived
loss of control increased probability of relapse
unable to cope with a high-risk situation resulting in the

use of cigarettes, experi
ences a conflict between his commitment to stop smoking

and his actions. This
conflict could result in dissonance, attributions to self,

and debilitating emotions.
As we have seen attributions to stable, internal causes

results in emotions like
guilt which have a negative impact on further task

behavior. In this situation that
will mean an increased probability of further relapses.

The absence of a coping response does not necessarily lead
to a lapse, it
"only" increases the probability of a lapse. The basis for

this assumption lies in
the relationship between the presence or absence of the

response and the increase
or decrease of the perception of one's own self-efficacy

in coping with the situa
tion. This implies that the better ingrained and more

automated the response is,
the higher one's self-efficacy and the lower the

probability of a relapse will be.
Take for instance the quitting smoker who needs to think

every time when he or
she is offered a cigarette whether or not he/she will

accept the cigarette compared
with the quitting smoker who automatically says: "No thank

you, I don't smoke
anymore." The pressures on the latter person will be much

smaller than on the
one who has to make up his/her mind anytime he or she is

offered the temptation
of a cigarette. One can conclude from the above-mentioned

theories that it is not enough to
motivate people to adopt healthier behaviors; one has to

equip them with the
necessary coping-skills to avoid the undesired behavior.

This will decrease the
probability of a relapse because it increases their

self-efficacy. If a lapse occurs
it is important to make sure that it is attributed to the

"right" cause. This means
that it should not be attributed to internal, stable

causes like willpower or ability,
as these attributions result in a lowered self-efficacy and

a lower expectation of
success which will result in a higher probability of a

relapse. In the last part of
this contribution we will discuss some examples of
interventions to avoid the pit
falls mentioned. One important source of self-efficacy

information is one's own experience
(Bandura, 1986). This implies that self-efficacy can be

raised by letting people
experience success. Marlatt and Gordon (1985) devised a

procedure, consisting
of four different stages, to influence the self-efficacy of

their clients. These
stages are: (a) influencing the frame of reference of the

client, (b) searching for
high-risk situations and the learning of coping-skills,

(c) actual practice of these
coping-skills, and (d) learning how to handle relapses.

The first stage consists of teaching the client that to
achieve the desired
behavior one has to learn certain skills. To quit smoking

for instance is not a
question of abilities or willpower but a question of

skills. This replaces ideas like
"I am a weakling who cannot quit smoking by himself' by the

idea that one is
going to learn the skills to quit smoking. As

self-efficacy is related to experienc
es of success, it is important to state realistic

subgoals. One such subgoal could
be to stop smoking during the day. The second stage

consists of looking at all the
possible high-risk situations, by means of keeping a diary,

self-monitoring or
exploring the reasons why previous quitting attempts

failed. For each high-risk
situation a coping response should be devised. Thirdly, it

is important to practice
these responses. A possibility to practice lies in actively

seeking the risk situa
tions and discussing the experiences afterwards. If a lapse

occurs, it is important
to learn from this lapse by examining what caused it and

how this can be avoided
in the future. Reattribution studies showed that not all

the causes are as good:
Attributing to ability has different results from

attributing to the situation.
Relapse prevention theory states that attributing to the

self is not a good attribu
tion because it will lead to negative emotions and a

lowered self-efficacy. One
has to attribute to external, unstable and controllable

aspects of the situation.
This recommendation is at odds with that of reattribution

research in which it is
recommended to attribute to internal, unstable and

controllable causes
(specifically effort) after failure. One could say that the

attribution recommended
in relapse prevention consist of finding a (better) coping

response for the situa
tion in which one failed. The cause of the lapse is

external, but one is responsible
for finding a way to handle that specific Cause and that

is something one has to
do by oneself. Relapse prevention and attribution theory do

agree, however, on
the subject of which attributions are not helpful after a

lapse; stable or uncon
trollable Causes whether they are internal or external will
result in a lowered self
efficacy or expectation of success and thus in a higher

probability of a total
relapse. Hall, Rugg, Turnstall, and Jones (1984) have

shown that such a coping-skills
treatment can be successful. They recruited 135 smokers who

followed an
aversive-smoking program. For half of them this was

followed by training in
coping skills. The other half participated in a discussion

group. Skills training
was more effective in avoiding relapse than the discussion

group. Analyses
showed that this training was especially effective with

subjects who smoked less
than 20 cigarettes a day. Killen, McCoby, and Taylor (1984)

also showed the
effectiveness of skills-training. Fifty-four subjects were

divided aCross three
conditions; nicotine chewing gum, skills-training, and a

combination of both.
Follow-up measures after 10 months showed abstinence of

smoking of respec
tively 23%, 30% and 50%. A combination of interventions waS

clearly the most
effective, but skills-training alone was more effective

than nicotine chewing gum. CONCLUSIONS Self-efficacy has
been shown to be an important determinant of behavior and
of behavior change. Self-efficacy expectations predict

future success or failure in
behavior change programs, like quitting smoking. Health

education should focus
on the improvement of self-efficacy in addition to

motivating people to behave in
a healthier way. Convincing people about the utility of the

expected behavior is
necessary but not sufficient. Especially people who have

experienced multiple
failures and interpret these failures as caused by stable,

internal and uncontrol
lable causes should undergo some kind of reorientation

period in which they are
taught that they are indeed able to change their behavior.

Furthermore we may
say that the relapse prevention model can be described in

attributional and self
efficacy concepts. This model states that it is not only

important to convince
people to adopt healthier habits and change their

attributions but they have to be
taught how to cope with difficult situations, that will

surely arise, as well. Some
of these difficulties will be caused by barriers which can

be overcome by the per
son concerned. These difficulties can be dealt with within

the theoretical-self
efficacy and attributional-framework, and the resulting

relapse prevention inter
ventions. Other difficulties will be caused by real

barriers (Bandura, 1986). The
difference between these two kinds of barriers is not

unequivocally determined:
Perhaps it is best to state that this difference lies in

the perceived controllability
of the barriers. Real barriers are beyond the person's

control; others might be
controllable. Next to skills training, health educators
should focus on health promotion
strategies in which barriers beyond the individual's

control, but controlled by
government or community, are removed as well (De Leeuw,

1989; see also
Bandura, 1986, p. 449, about collective efficacy and social

action). Most of the
studies that were presented focussed on smoking, but the

basic ideas can be gene
ralized to other areas in health education, for instance

AIDS preventive behavior
Ajzen, I. (1987). Attitudes, traits and actions:

Dispositional prediction of behavior in
personality and social psychology. In L. Berkowitz (Ed.),

Advances in experimental
social psychology, (Vol. 20, pp. 1-63). New York: Academic

Press.
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes

and predicting social behav
Anderson, C. A., Horowitz, L., & French, R. (1983).

Attributional style of lonely and
depressed people. Journal of Personality and Social

Andrews, G. R., & Debus, R. L. (1978). Persistence and the

causal perception of failure:
Modifying cognitive attributions. Journal of Educational


Prentice Hall.

promoting behavior. In 1. Madden (Ed.), Neurobiology of

Clifford, M. M. (1986). The effects of ability, strategy

and effort attributions for educa
tional, business and athletic failure. British Journal of

Educational Psychology, 56,
169-179.
Colletti, G., Supnick, 1. A., & Payne, T. 1. (1985). The

smoking self-efficacy question
naire (SSEQ): Preliminary scale development and validation.

Behavioral Assessment,
7,249-260.
De Leeuw, E. (1989). The sane revolution. Health promotion:

Backgrounds, scope,
prospects. Maastricht, The Netherlands: Van Gorcum.
Den Boer, D. 1., Meertens, R. M., Kok, G. J., & Van

Knippenberg, A. (1989). Measure
ment effects in reattribution research. European Journal

of Social Psychology, 19,
553-559.
Den Boer, D. J., Meertens, R. M., Kok, G. J., & Van

Knippenberg, A. (1990). Weiner's
attribution theory: A test of causal relations. Internal

report, University of Limburg,
Maastricht, The Netherlands.
De Vries, H. (1989). Smoking prevention in Dutch

adolescents. Dissertation, University
of Limburg, Maastricht, The Netherlands.


De Vries, H., & Kok, G. J. (1986). From determinants of

smoking behavior to the impli
cations for a prevention programme. Health Education

Research, 1, 85-94.
Eiser, J. R., & Van der Pligt, J. (1986). Smoking cessation

and smokers' perceptions of
their addiction. Journal of Social and Clinical Psychology,

37, 261-272.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude,

intention and behavior: An introduc
tion to theory and research. Reading, MA: Addison Wesley.
Forsterling, F. (1988). Attribution theory in clinical

psychology. New York: John Wiley
and Sons.
Fowler, J. W., & Peterson, P. L. (1981). Increasing reading

persistence and altering
attributional style of learned helpless children. Journal

of Educational Psychology,
73,251-260.
Green, L. W., & Lewis, F. M., (1986). Measurement and

evaluation in health education
and health promotion. Palo Alto, CA: Mayfield.
Hall, S. H., Rugg, D., Turnstall, C., & Jones, R. T.

(1984). Preventing relapse in ciga
rette smoking by behavioral skill training. Journal of

Consulting and Clinical
Heider, F. (1958). The psychology of interpersonal
relations. New York: John Wiley
and Sons.
Hewstone, M., & Antaki, C. (1988). Attribution theory and

social explanations. In
Hewstone, M., Stroebe, W., Codol, 1. P., & Stephenson, G.

M. (Eds.). Introduction to
social psychology. A European perspective (pp 111-14]).

Oxford: Blackwell.
Hospers, H. J., Kok, G. J., & Strecher, V. J. (1990).

Attributions for previous failures
and subsequent outcomes in a weight reduction program.

Health Education
Quarterly, 17,409-415.
Janoff-Bulman, R. (1979). Characterological versus

behavioral self-blame: Inquiries
into depression and rape. Journal of Personality and

Social Psychology, 37, 1798
1809.
Jonkers, R., De Haes, W. F. M., Kok, G. J., Liedekerken,

P., & Saan, J. A. M. (1988).
Effektiviteit van GVO (Effectiveness of health education).

The Hague, The Nether
lands: Care.
Kelley, H. H. (1967). Attribution theory in social

psychology. In Levine, D. (Ed.)
Nebraska symposium on motivation (Vol. ]5, pp 192-240).

Lincoln: University of
Nebraska Press.
Killen, D., McCoby, N., & Taylor, C. B. (1984). Nicotine

gum and self-regulation train
ing in smoking relapse-prevention. Behavioral Therapy,

15,234-248.
Kirsch, I. (1985). Self-efficacy and expectancy: Old wine

with new labels. Journal of
Kirsch, I. (1986). Early research on self-efficacy: What we

already know without
knowing we knew. Journal of Social and Clinical

Psychology, 4(3), 339-358.
Kok, G. J. (1988). Health motivation: Health education from

a social psychological
point of view. In S. Maes, C. D. Spielberger, P. B.

Defares, & I. G. Sarasen (Eds.),
Topics in health psychology (pp. 295-300). New York: Wiley.
MarIan, G. A., & Gordon, 1. R. (1985). Relapse prevention.

Maintenance strategies in
the treatment of addictive behaviors. New York: The

Guilford Press.
McHugh, M., Beckman, L,. & Frieze, I. H. (1979). Analyzing

alcoholism. In I. H.
Frieze, D. Bar-Tall, & 1. Carrol (Eds.). New approaches to

social problems (pp 168
208). San Francisco: Jossey Bass.
Mudde, A. M., Kok, G. J., & Strecher, V. 1. (1989, July).

Self-efficacy and success
expectancy as predictors of the cessation of smoking.

Paper presented at the First
European Congress of Psychology, Amsterdam, The

Netherlands.
Schunk, D. H. (1984). Sequential attributional feedback and

children's achievement
behaviors. Journal of Educational Psychology, 76,

1159-1169.
Schwartz, J. L. (1987). Review and evaluation of smoking
cessation methods: The
United States and Canada. 1978-1985. U.S. Department of

Health and Human
Services, Public Health Service. Bethesda, MD: National

Institutes of Health.
Strecher, V. 1., DeVellis, B. M., Beker, M. H., &

Rosenstock, I. M. (1986). The role of
self-efficacy in achieving health behavior change. Health

Education Quarterly, 13,
73-91.
Weary, G., Stanley, M. A., & Harvey, J. H. (1989).

Attribution. New York: Springer.
Weiner, B. (1985). An attributional theory of achievement

motivation and emotion.
Weiner, B. (1986). An attributional theory of motivation

and emotion. New York:
Springer.
Wilson, T. D., & Linville, P. W. (1982). Improving the

academic performance of
college freshman: Attribution therapy revisited. Journal of

Psychology, 42,367-376. THE INFLUENCE OF EXPECTANCIES

AND PROBLEM-SOLVING STRATEGIES ON SMOKING INTENTIONS
Martin V. Covington and Carol L. Omelich Educational
researchers have proposed that children's health risktaking
behavior depends in part on how youngsters resolve
social/interpersonal problems and on the personal
expectations that arise in the course of their daily
lives. The present research attempted to establish the
usefulness of such a problem-solving/social-cognitive
approach to anti-smoking interventions by (a)
investigating the nature and causes of problem-solving
deficiencies among students at high risk for cigarette
smoking, and (b) by determining if various selfefficacy
problem-solving elements influence smoking decisions so as
to alter the level of temptation experienced in social
situations. Some 4,000 sixth, eighth, and tenth graders
(/1and 15-year-olds) reacted to hypothetical smoking
scenarios by rating temptation level, intentions to smoke,
and the likelihood of applying ten problem-solving
strategies thought to mediate smoking intentions.
Between-group analysis indicated consistent differences
in problem-solving approaches among individuals with
various smoking histories (e.g., nonsmokers, regular
smokers), regardless of grade, sex, ethnicity or ability
level. Moreover, path analysis showed that self-efficacy
strategies and outcome expectations acted as mediators of
the temptation -. intention relationship regardless of
smoking history. Implications of these results for a
social problem-solving approach to adolescent health
education are considered.
Effective problem solving has long been a valued teaching

goal in schools. In
more recent years, social problem resolution has also come

to occupy a central
role in thinking about how to achieve the larger

objectives of health education
including long-term risk reduction and health maintenance,

especially in the area
of cigarette use (Botvin & Eng, 1980, 1982; Botvin, Eng, &
Williams, 1982;
Cohen et al., 1988; Gilchrist, Schinke, & Blythe, 1979;

Kim, 1982; Schinke,
Gilchrist, Snow, & Schilling, 1985; Williams & Arnold,

1980). UsualJy the
working definition of a problem and its solution involves

the resolution of an
interpersonal conflict between family members or among

peers in which health
issues are not necessarily the paramount concern of the

participants, but rather
get resolved in the course of solving certain other issues

involving children's
needs for autonomy, recognition and affiliation. According
to this view, anti-smoking education should provide a range
of
problem-solving skills for coping with larger issues of

personal/social signifi
cance, yet in a manner that also reduces the likelihood of

problem resolution in
favor of smoking. It is in this sense that smoking

behavior should not be the sole,
or even the primary, concern of anti-smoking

interventions. If basic needs such
as those for autonomy and affiliation underlie adolescent

smoking decisions,
educational interventions must work within the context of

these naturally occur
ring motivations and not against them for short-term

prevention of smoking.
This growing concern with the problem-solving dynamics

involved in health-risk
reduction is further reflected in the concept of informed

decision making. The
operative notion here is that students should be taught

the judgmental skills
necessary to reach their own conclusions regarding

personal, moral and social
issues, thereby shifting the burden of defining the

content of personally relevant
education from the school and teacher to the student

(Botvin, 1983; Botvin et al.,
1980; Jones, Piper, & Matthews, 1970; Piper, Jones, &

Matthews, 1974). However, despite the widespread
recognition of the potential health rele
vance of various problem-solving models, few health

intervention programs are
based primarily on social decision making (for a review,
see D'Onofrio, 1983b).
One of the reasons for this dearth, at least in the area

of cigarette smoking, is the
general absence of information on the problem-solving

characteristics of smokers
and nonsmokers. Missing is systematic evidence on the

question of whether or
not smokers differ from nonsmokers in how they approach a

smoking decision;
and how such differences, if they exist, enter into smoking

decisions. Yet the in
herent value of a problem-solving model to the goals of

health education depends
fundamentally on the presumption of such skill

differences, and on their causal
role as mediators of a smoking decision. For example, it

may be that smokers ar
rive at different smoking decisions than do nonsmokers, not

because they see the
issues differently or choose different strategies to

resolve interpersonal dilemmas,
but simply because their peer culture encourages smoking or

because their par
ents condone it. If this is true, a problem-solving

approach offers little leverage
for an increased understanding of smoking behavior for its

interdiction. In this
instance, educators would be better advised to direct their

attention to other
aspects of the process of smoking uptake and resistance,

perhaps by focusing on
parent education or by supporting more restrictive

legislation on the use and
availability of cigarettes among adolescents. Obviously,
then, an important initial step in establishing the
usefulness of a
problem-solving focus in health education is to

investigate the nature and causes
of problem-solving deficiencies among adolescents at risk

for smoking. Such an
investigation was the main purpose of the present study.

Several potential sources of problem-solving differences
can be anticipated,
each of which may lead to a decision to smoke or not, and

each implying some
what different intervention strategies. First, it may be

that individual differences
in the perception of smoking issues and in the availability

of prosocial strategies
are conditioned in part by variations in basic ability.

Although many smokers are
bright, well-informed individuals, there are enough

scattered reports indicating
that children who smoke tend to score lower on standard

achievement tests and to
have poorer scholastic records to make this hypothesis of

basic cognitive differ
ences between smokers and nonsmokers more than idle

speculation (for a review,
see Evans, Henderson, Hill, & Raines, 1979). Intervention

strategies would
likely differ markedly depending on the acceptance or

rejection of this
possibility. A second and likely more important source of

differences involves the indi
vidual's personal smoking history. Smokers and nonsmokers

face fundamentally
different decisions when confronted with the possibility of
smoking. Non
smokers feel considerable pressure to yield arising from

personal curiosity and
from peer pressure (Covington & Omelich, 1986), yet by

definition are highly
resistant to such temptation. Such resistance likely

depends in part on a well
developed repertoire of problem-solving strategies which

acts to diffuse situa
tions in favor of not smoking. By comparison, the

experienced smoker tends to
yield more easily to temptation, irrespective of peer

pressure and immediate cir
cumstances, and sees little violation of his or her

self-image by smoking. As a
result, young smokers are likely to be less concerned with

the issues surrounding
smoking and with the probable consequences of their

behavior, and to possess
less well-formulated strategies to avoid smoking. Such

concerns may simply be
less salient to smokers, not only because the physical

addictive process may have
advanced to a point so as to render the decision to smoke

automatic (Leventhal &
Cleary, 1980), but also because young smokers possess

well-entrenched rationa
lizations that minimize and distort the risks of having

"just one more cigarette."
Thus, unlike the calculated response of the nonsmoker to

peer pressure, what is
likely to dominate the decision for smokers is their

perception of the value of cig
arettes to achieve certain desired goals, the force of

habit, and the whim and
moods of the moment. This line of reasoning suggests

that we should be quite surprised if investi
gators failed to find differences in the characteristic

ways smokers and nonsmok
ers typically approach a smoking decision. However, if it

is so obvious as to why
such differences should arise, then perhaps their existence

is merely trivial from
the perspective of intervention. In effect, we must also

address a further ques
tion: Do such problem-solving differences actually exert a

causal impact on im
portant target behaviors such as one's intentions to begin

smoking or to continue
to smoke in the future? As we have argued, problem-solving

differences may
simply be an artifact of the individual's smoking

experiences-more the result of
smoking or of coping in a smoking milieu than a cause of

smoking. These causal questions are considered from a
social-cognitive perspective
which emphasizes the role of expectations as leading causes

of behavioral inten
tions (Bandura, 1977, 1989). According to social-cognitive

theory at least two
kinds of expectations can be discerned: First, outcome

expectancies which refer
to the perceived consequences of one's actions and their

importance to the indi
vidual; and second, self-efficacy expectations, namely,

perceiving that one is cap
able of performing a specific action. Self-efficacy has
proven an especially
robust predictor of intentions to smoke or to abstain, even

after other factors such
as attitudes toward smoking have been taken into account

(Ajzen & Madden,
1986; De Vries, Dijkstra & Kuhlman, 1988). Moreover,

individuals report great
er feelings of efficacy after episodes in which they

resisted smoking than when
they succumbed to the temptation to smoke (Garcia, Schmitz,

& Doerfler, 1990).
For purposes of this research we have focused on two

aspects of outcome expect
ancies. First, there are the outcomes themselves, that is,

the anticipated
consequences of, say, one's parents finding out that a

child has been smoking.
Then, second, there is the importance the individual

attaches to various out
comes, or what we refer to as issues: To extend our

example, how important the
anticipated parental reactions are to the child. Issues and

consequences often
interact in their influence on intentions. For example, the

expectation of a swift
reprisal from parents may have little to do with the

child's decision to smoke or
not if he or she does not care what the parents think or

do. Additionally, we have
represented the self-efficacy portion of Bandura's model as

the perceived avail
ability of strategies for effective action such as

diffusing conflict and negotiating
a positive, nonsmoker resolution which might include

communicating one's
values to others and redirecting attention from smoking to

other more con
structive activities. These three components-perceived

issues, consequences, and strategies
comprised the focus of this investigation. In order to

evaluate the saliency of
these three components to a smoking decision, we focused on

the pervasive
smoking temptation/intention linkage. It has been well

established that the temp
tation value of a cigarette exerts a dominant, if not

preemptive, influence on
intentions to smoke among young adolescents (Best &

Hakstian, 1978; Coving
ton & Omelich, 1988), in effect, as temptation (desire)

increases, so do intentions
(desired action). For the child without a well-developed

set of internalized con
trols, desire typically translates directly into action).

Research on the develop
ment of moral and social reasoning suggests that such

impetuousness only slowly
comes under the control of evolving cognitive processes and

self-interested
reflection. Unfortunately, such self-regulating cognitions

are not sufficiently
established until most youngsters are well past junior high

school (13-15 years
old), which appears to be the most vulnerable periods to

smoking temptation
(Covington & Omelich, 1986). If we accept temptation level
as a pivotal causal
factor in smoking uptake, then a reasonable measure of the

value of a problem
solving approach to moderating health risks would be

reflected to the extent to
which cognitive factors act to offset the dependency of

intentions on temptation
so that intentions to smoke will remain low, irrespective

of the temptation of a
given circumstance. This reasoning implies a causal model

of the kind portrayed
in Figure 1. The unidirectional arrows indicate the paths

of influence that were
examined in the present study. Put in causal terms, we

asked whether or not
present problem-solving skills will reduce the dependency

of intentions on temp
tation: (a) by serving as an inhibitor of temptation (Le.,

high tempta
tion -+ problem-solving strategies -+ low intentions);

and/or (b) by providing a
direct, countervailing presence in their own right

(problem-solving
strategies -+ low intentions)? Smoking Temptation

Health Issues R2 = .047 Decision-Making Strategies R2
= .033 Interpersonal Strategies R2 = .086 Self-Peer
Strategies R2 = .038 .440 Affective Consequences R2 =
.104 Needs Consequences R 2 = .053 Intention to Smoke
R 2 = .406
Figure I Path diagram of the hypothesized effect of problem

solving in moderating the temptation/intention
relationship in smoking decisions. . 2 1 8 . 2 1 8 . 2 9
4 . 2 1 8 . 0 8 9 . 0 8 9 . 0 4 9 . 0 9 6 . 0 9 6 . 0 4 8 .
3 2 2 . 3 2 2 The history of research on problem solving
is as rich and variegated as it is
voluminous (for a review, see Covington, 1986; Hermstein,
Nickerson, Sanchez,
& Swets, 1986; Nickerson, Perkins, & Smith, 1985). Surely,

some approaches
will prove more useful to the goals of health education

than will others, but
which ones? While issue, consequence and strategy

components may all contrib
ute to a decision to smoke or to resist, their impact is

unlikely to be of equal
weight across all individuals. Thus in order to provide a

firm empirical basis for
problem-oriented interventions, we must also know which of

these components
are most influential in controlling the smoking intentions

and behavior of
youngsters of differing ages, sex, intellectual ability and

ethnic group member
ship. Most important, key problem-resolution components may

also differ
depending on the youngster's personal smoking history. In

effect, the cognitive
considerations that influence the confirmed smoker's

decision to smoke another
cigarette may be quite different from the factors that

enter into the nonsmoker's
decision about whether or not to begin smoking at all.

Likewise, the dynamics of
both of these groups are likely to differ in tum from those

of youngsters who
have experimented previously with cigarettes, but who have

yet to continue the
practice. Given the above considerations, the specific

purpose of the present study was
twofold: (a) to determine if different expectancies and
problem-solving ap
proaches that children characteristically bring to a

smoking decision vary as a
function of the individual's smoking history, age, sex,

intellectual ability, and
ethnic group membership; and (b) to determine if any

variations in these
problem-solving elements influence smoking intentions in

such a way as to alter
the temptation value of cigarettes, and whether or not

these causal dynamics
differ depending on smoking-group membership. METHODS

Subjects and Procedures The data reported in the present
study were gathered in connection with the
research and development of a school-based intervention

program by the Risk
and Youth: Smoking Project (RAY:S) at the University of

California at Berkeley
(Covington, D'Onofrio, Thier, Schnur, & Omelich, 1983;

D'Onofrio, Thier,
Schnur, Buchanan, & Omelich, 1982). The project consisted

of three inter
locking phases: (a) a program of basic research and theory

building; (b) the
practical implication of these research findings in the

development of educational
intervention techniques; and (c) the field testing and

formal evaluation of these
instructional products both in schools and informal

community settings. As part of the research program, a
smoking risk profIle questionnaire was
administered to 6,494 students in 51 schools in 12 cities

in the great San
Francisco Bay Area. Of these students, 49.2% (3,198) were
sixth-graders (I1-12
years old); 31.2% (2,025) were eighth-graders (13-14

years); and 19.6% (1,271)
were in the tenth grade (15-16 years). The total sample

comprised 8.7% Asians,
32.0% Blacks, 8.6% Hispanics, 48.6% Whites, and 2.1 % who

classified them
selves as "Other." This ethnic distribution roughly

approximated the population
of the San Francisco Bay Area. The 179-item questionnaire

was designed to identify the prevalence of smok
ing and the causes of cigarette uptake, resistance and

cessation. The predictor
variables included distal antecedents (e.g., family

characteristics), proximal ante
cedents (e.g., peer-group characteristics), beliefs

regarding the instrumental value
of cigarette use, and the quality of prosocial

decision-making skills. One section of the questionnaire
presented several brief, life-like scenarios in
which youngsters might be tempted to smoke. Each situation

depicted tempta
tion in a group setting varied along several dimensions

including composition of
the group (e.g., best friends versus casual acquaintances)

and whether or not
others in the group were already smoking. Because the

reactions of subjects dif
fering in degree of previous experience with cigarettes

have been shown to be
little influenced by these variations in circumstances

(Covington & Omelich,
1988), all situational variables of interest were combined
across scenarios for the
present analysis. Criterion Measures All subjects

responded to a series of Likert-type rating questions as
follows: Temptation. "In this situation, how much would
you actually want to
smoke?" (1 = not at all; 5 = very much) Intentions. "Do

you think you would actually smoke in this situation?" (l =
definitely no; 5 = definitely yes). Problem-solving. Ten

theoretically derived scales incorporating 28 items
represented the basic problem-solving components thought

to mediate smoking
intentions (Covington, 1981; Fishbein, 1977; Sutton, 1989).

As a group, these
ten scales measured: (1) the extent to which youngsters

were likely to consider
various issues as important in making a smoking choice (3

scales); (2) the degree
to which they were likely to entertain various prosocial

strategies facilitative of a
nonsmoking decision (4 scales); and (3) the extent to

which children anticipated
various consequences following a decision to smoke (3

scales). Preliminary prin
cipal components analyses of the total sample provided

empirical validation of
the conceptual structure of the 28 items selected to

measure these ten problem
solving scales. The factor solutions accounted for 55.9%

of the variance in the
item set concerned with issues, 47.5% for strategies, and

43.0% for conse
quences. The actual items making up the various scales are

available upon
request from the authors. Issues. Three Likert-type scales
measured the perceived importance of
affective issues, needs issues and health issues,

respectively (item responses: 1 =
not important; 4 = very important). The affective issues

scale (range: 3-12)
measured the importance of anticipated affective reactions

of one's parents and
peers, and self-reactions to a decision to smoke (M =

9.15, SD = 2.03, a. = .48).
The needs issues scale (range: 4-16) assessed the extent

to which the subject
considered needs for affiliation, for appearing mature, and

for a sense of inde
pendence as important issues in making a decision about

smoking) (M = 9.36,
SD = 2.50, a. = .51). The health issues scale (range: 2-8)

measured whether or
not concerns for potential health risk and addiction were

important issues in a
smoking decision (M = 6.93, SD = 1.51, a. = .69).

Strategies. Four Likert-type sales appraised whether or not
subjects would
utilize a variety of prosocial, problem-solving skills

(item responses: 1 =
definitely no; 5 = definitely yes). A decision-making

strategy scale (range: 2-10)
measured whether or not subjects were inclined to reflect

on a smoking decision
rather than to act impulsively without deliberating (M =

7.08, SD = 2.61, a. =
.76). An interpersonal strategy scale (range: 5-25)

appraised the subject's
potential repertoire of interpersonal skills for defusing

pressures to smoke (M =
16.99, SD = 4.86, a. = .76). A parent strategy scale

(range: 2-10) assessed the
degree to which subjects would be likely to seek out and

consider parental
opinion (M = 6.73, SD = 2.29, a. = .61). A self/peer

strategy scale (range: 2-10)
reflected a preoccupation with satisfying one's own wishes

or those of one's
peers in reaching a decision (M = 4.77, SD = 1.92, a. =

.21). Consequences. Three Likert-type scales appraised the
extent to which sub
jects anticipated various consequences should they decide

to smoke (item
responses: 1 = not at all; 5 = very much). The affective

consequences scale
(range: 3-15) measured the expected degree of parental,

peer, and self
displeasure should the subject smoke (M = 10.82, SD =

3.06, a. = .61). The needs
consequences (scale range: 3-15) indicated whether or not

smoking would pro
mote a sense of autonomy, maturity, and belonging for the

subject (M = 7.18,
SD = 2.78, a. = .37). The health consequences scale (range:

2-10) assessed the
degree to which subjects anticipated health problems as a

result of smoking (M =
7.37, SD = 2.29, a. = .38). Individual Difference Measures

Smoking status. Current smoking status was measured by a
single, self
report item as follows: "Check the one sentence below that

best tells about you":
"I have never smoked a cigarette (not even a few puffs)";

"J tried one or two cig
arettes and never smoked again"; "I smoked for a while,

but do not smoke any
more"; "I smoke cigarettes regularly." The incidence of

cigarette use in the final
sample of 3,994 students having a complete data file

closely paralleled that found
in national self-report surveys of adolescents in the

United States (National
Institute of Education, 1979). Some 1,804 subjects had

never smoked (NS) (N =
1,127, N = 460, and N = 217 at grades 6, 8, and 10,

respectively). Among the
remaining subjects, three additional groups were

differentiated. That group
which had smoked only one or two cigarettes (Experimental

Smokers; ES); that
which had smoked for a while and then quit (Ex-Smokers;

XS); and that which
smoked regularly at the time of assessment (Regular

Smokers; RS). There were
1,256 ES subjects (N = 518, N = 424, N = 314, at grades 6,

8 and 10, respec
tively); 433 XS subjects (N = 125, N = 180, and N = 128,

respectively); and 510
RS subjects (N = 89, N = 235, and N = 177, respectively).

While the validity of adolescent self-reports of smoking
has been questioned
repeatedly, a study by Wi11iams, Eng, Botvin, HiIJ, and

Wynder (1979) demon
strated that when assured of anonymity (as was the case in

the present research),
adolescents do give accurate self-reports of smoking

behavior. AdditionaUy, by
applying a bogus-pipeline procedure (Evans, Hansen, &
Mittlemark, 1977) to a
subsample (N = 193) of our sixth-grade and eighth-grade

subjects, we also con
fmned the validity of self-report measures. A comparison of

this subsample with
the main sample without bogus pipeline revealed no

significant differences in re
ported smoking status, p < .05. The reported smoking

levels for the main sample
and the bogus pipeline subsample were comparable: NS (50.7%

versus 54.0%),
ES (29.5% versus 27.0%), XS (9.5% versus 9.0%), RS (10.3%

versus 10.0%). Sex and grade. Sampling was
disproportionately heavier at the lower grades
(11-14 years old) to allow for adequate numbers of

youngsters with smoking ex
perience (N6 = 1,859, N g = 1,299, NIO = 836). Boys (N =

1,949) and girls (N =
2,045) were represented in roughly equal proportions across

the three grade
levels. Ability. Academic ability was measured by an

abbreviated IO-item form of
the Comprehensive Test of Basic Skil1s (CTBS/McGraw Hill,

1973). Three
forms, one appropriate to each grade level, were used so

that items of increasing
difficulty insured maximum discrimination across age.

Ethnicity. Subjects were also classified as to ethnic
background based on
their response to a single, six-category questionnaire

item: Asian, Black,
Hispanic, Latino, White, and Other. Since the Latino

category was used by less
than 1 % of the sample, it was combined with the Hispanic
category. Also, due to
the undifferentiated nature of the "Other" category, these

subjects (2.1 %) were
dropped from those analyses involving ethnicity as a

factor. Statistical Analysis Between-group differences in
smoking temptation level, smoking intentions,
and all problem-solving variables were assessed by separate

3 (Grade) x 2 (Sex)
x 2 (Ability) x 4 (Ethnicity) analyses of variance.

Preliminary analysis indicated
that the inclusion of Smoking Status in a five-way

factorial design produced too
many depleted cells and resulted in no significant

interactions with the other indi
vidual difference factors. Accordingly, analysis of

smoking-status group differ
ences were investigated through separate one-way analyses

of variance. Deter
mination of the specific source of any significant group

differences involved the
use of Dunn's multiple-comparison procedures (Kirk, 1968)

which permit testing
a specific number of a priori contrasts at a predetermined

level of significance
(a = .05). The postulated causal role of the various

problem-solving elements
and temptation in determining smoking intentionality were

evaluated by a path
analytic interpretation of multiple regression. Path

analysis (Pedhazur, 1982)
allows for all determining factors as specified by a

causal model to be incorporat
ed into an overall predictive analysis, thereby permitting
an estimation of the
relative contribution (both direct and indirect) of each

determinant to variations
in dependent variables of interest. Hypotheses regarding

the similarity of these
causal relationships for subjects with varying smoking

histories were assessed by
comparing the differences in magnitude of regression slopes

for the relevant
groups. RESULTS AND DISCUSSION Between-Group Differences

Table 1 displays the mean values and standard deviations
for smoking inten
tions and for the ten issue, strategy, and consequence

variables by smoking status
and grade. A 2 (Sex) x 3 (Grade) x 2 (Ability) x 4

(Ethnicity) analysis of var
iance was performed on each of the variables shown in

Table 1. With the excep
tion of grade level, the impact of these individual

difference variables-although
significant due to the large sample involved-was marginal

in terms of magni
tude of explained variance «(02). For this reason, these

mean values are not
tabled, nor are those for temptation level since these

latter results are essentially
identical to those found for intentions (r = .57, P <

.05). Ability. Brighter youngsters found each of the
smoking issues, strategies, and
consequences, with the exception of the self/peer strategy

variables (NS), to be
more salient to their thinking about a smoking decision

than did less bright
youngsters, all Fs (1,3946) > 3.84, p < .05. However, the
amount of explained
variance «(02) accounted for by ability in each dependent

factor was negligible,
accounting on average for less than 1 % of the variance.

The sole exception was
the health consequences variable, for which ability level

explained some 4% of
the variance, F(l,3946) = 122.79, (0 2 = .044. It appears

that brighter youngsters
weigh the health consequences of their actions more

heavily (N = 7.87) than do
less bright individuals (M = 6.93). Otherwise, however, we

conclude that
problem-solving dispositions are not particularly dependent

on differences in
basic cognitive ability. Nor did variations in ability

influence smoking tempta
tion level «(02 = .(00) or intentions to smoke «(02 =

.004), although the ability
factor proved significant for temptation and intentions,

Fs (1,3946) = 5.55,24.24,
respectively, p < .05, owing to the large sample size. One

caution should be
noted, however. The recognition mode of assessment used in

this study may
underrepresent the role of ability in real-life dilemma

resolutions. For example,
had our subjects been assessed under a free-response

format, then ideational flu
ency factors, verbal skills, and analytic reasoning would

have likely become
more salient, thereby increasing the dependency of

responses on ability level.
Table 1
Mean Values and Standard Deviations on Intentions to Smoke

and Problem-Solving
Skills by Current Smoking Status and Grade
Dependent Smoking Status Grade
Variable NS ES XS RS 6 8 10
(N) (1804) (1256) (433) (501) (1859) (1299) (836)
Intenuon~ M 1.32 1.72 2.10 3.26 1.63 1.98 1.80 SD

0.69 1.01 1.18 1.31 1.02 1.23 1.18
Issues
Affective M 9.55 9.06 8.67 7.97 9.35 8.99 8.66 SD 2.01

1.95 2.02 1.92 1.99 2.05 2.10
Needs M 9.34 a 9.34 a 9.37 a 9.39 a 9.51 9.28 9.06

SD 2.70 2.44 2.23 2.15 2.79 2.23 2.15
Health M 7.14 6.97 6.66 6.03 7.12 6.72 a 6.63 a SD

1.41 1.42 1.54 1.80 1.41 1.59 1.62
Strategie!!
DecisionM 7.40 7.08 6.60 5.88 7.21 6.91a 6.75 a
making SD 2.71 2.55 2.41 2.32 2.67 2.54 2.64
InterM 18.24 17.04 15.38 13.54 18.02 16.19 15.64
personal SD 4.62 4.56 4.61 4.45 4.71 4.68 4.88
Parent M 7.27 6.71 6.05 5.64 7.26 6.36 a 6.19 a SD

2.23 2.18 2.28 2.25 2.25 2.25 2.24
Self/peer M 4.39 4.89 b 5.13 b 5.44 a 4.47 5.00 a 5.05

a SD 1.93 1.91 1.79 1.78 1.95 1.87 1.86
Conse!':lYences
Affective M 11.82 10.65 9.91 8.17 11.42 10.41 9.90 SD

2.74 2.81 2.88 2.97 2.86 2.98 3.28
Needs M 6.92 7.23 a 7.33 a 7.93 7.23 7.34 a 6.89 SD

2.83 2.72 2.67 2.67 2.88 2.64 2.72
Health M 7.58 7.25 a 7. lOa 6.70 7.34ab 7.2Jb 7.41 a

SD 2.29 2.25 2.24 2.32 2.29 2.27 2.36
Note. Similarly superscripted values (e.g., a, b) are not

significantly different (p < .05). NS = Nonsmokers; ES =
Experimental Smokers; XS = Ex-Smokers; RS = Regular
Smokers. Sex and ethnicity. Like ability, both the sex and
ethnic effects were signifi
cant, all Fs < 3.84, .05 P < .05, but accounted for only
negligible portions of the
problem-solving issues, strategies, and consequence

factors. Typical 0)2 values
were less than .005, especially for the sex effect. One
exception was the self/peer
strategy category for which ethnic-group membership

accounted for approxi
mately 3% of the variance. Several other exceptions

occurred for ethnicity,
which explained some 2-3% of the variance in

problem-solving behavior. For
example, it appears that Asians and, to a lesser degree,

Whites, were more likely
than Blacks to reflect upon the issues involved in a

smoking decision (decision
making strategy). Moreover, the issues of greater

importance to Asians and
Whites were the negative reactions of others (affective

issues) and the potential
health hazards of smoking (health consequences). On the

other hand, Hispanics
and Blacks were less likely than Asians and Whites to be

guided by peer group
suggestions for resolving dilemmas (self/peer strategy).

Finally, planned pair
wise contrasts performed after the ethnicity effect proved
significant for tempta
tion level and intentions, F(3,3946) = 14.65, 18.93,

respectively, indicated that
Asians were less tempted to smoke and consequently had

lower intentions to do
so when compared to the other three ethnic categories, p <

.05. By contrast,
Whites, Blacks, and Hispanics did not differ among

themselves with regard to
degree of temptation and intentions to smoke, p < .05. In

summary, with only the occasional exceptions noted above,
variations in
ethnic group membership, ability, and sex proved to be

only marginal contrib
utors to characteristic approaches to smoking dilemmas. Nor

were there any sig
nificant interactions among these group differences that

accounted for more than
negligible proportions of variance, all 00 2 < .005.

Thus, regarding the relative
degree of problem-solving responsiveness, we conclude that

it makes little differ
ence if a youngster is male or female, Hispanic or Asian,

or bright or less bright. Grade level. Grade level proved
to be a consistent, although modest, contrib
utor to variations in the problem-solving components, as

well as variations in
temptation level and intentions, all Fs(2,3946) < 3.00, p <

.05. The significance
of all pairwise contrasts is indicated by superscripts in

Table l. Similarly, super
scripted values were not significantly different from one

another (p < .05). The
results of pairwise contrasts generally revealed that the
younger the child, the
more salient were issues concerning affective reactions (00

2 = .020) and health
(00 2 = .026), and the more likely he or she was to

recognize the value of inter
personal strategies for problem resolution (00 2 = .044)

and to seek parental advice
(parent strategy: 00 2 = .044) while rejecting peer

suggestions and the paths of
personal rebelliousness (self/peer strategy: 00 2 = .020).

Without such con
straints, older students were more likely to be tempted by

cigarettes (00 2 = .017)
and to harbor stronger smoking intentions (00 2 = .023).

These age trends were
especially pronounced among eighth-grade students who

experienced greater
temptation and intentionality than either sixthor

tenth-graders. Smoking status. The single overwhelming
source of individual differences in
the saliency of problem-solving components was the current

smoking status of
the individual (Table 1). Smoking status accounted for some

30% of the varia
tion in intentions to smoke (FI3,39461 = 704.73,p < .05;

00 2 = .299. A series of
pairwise contrasts was performed between all smoking-status

groups for each
problem-solving factor separately and for both smoking

intentions and tempta
tion as well. Not unexpectedly, the four smoking-status

groups were all signifi
cantly differentiated as to intentions (as seen by the

absence of superscripts), with
current smokers reporting greatest degree of

intentionality and nonsmokers re
porting the lowest (RS > ES > NS, p < .05). Feelings of
temptation also depended
largely on one's smoking history, F(3,3946) = 426.96, P <

.05; 00 2 = .219, with
the same significant rank ordering by smoking-status

category for temptation as
was found for intentions. Also, as expected, with the

exception of the needs issues category (NS, as
indicated by the "A" superscript in Table 1), the smoking

status main effect for
all problem-solving issues, strategies, and consequences

factors proved signifi
cant, all Fs(3,3946) > 2.60, p < .05, .016 < 00 2 < .156.
A series of pairwise
contrasts revealed a consistent profile among individuals

with different smoking
histories. To summarize the data reported in Table t: With

regard to issues,
nonsmokers considered health matters and the opinions of

others (affective
issues) to be of greater relevance in decision making than

did all other smoking
groups, with health and affective concerns being least

salient among regular
smokers. Likewise, regarding perceived consequences,

nonsmokers were more
likely than were all other groups to view cigarette use as

a violation of both their
self-interest and their expectations of the reactions of

others (affective conse
quences), thereby anticipating greater emotional upset if
they smoked. Addition
ally, nonsmokers perceived the use of cigarettes as less

likely to result in feelings
of maturity and autonomy (needs consequences) and to create

greater health haz
ards than did all other groups (health consequences).

Finally, regarding strate
gies, nonsmokers were more likely to adopt a thoughtful

decision-making mode
than were any other groups, with such reflectivity being

least in evidence among
regular smokers. Moreover, interpersonal strategies

involved in the processes of
negotiation were more salient among nonsmokers, as well as

a willingness to
discuss such tactics with their parents (parent strategy).

Conversely, regular
smokers displayed a greater tendency to follow the advice

of their friends and/or
respond to their own desires (self/peer strategy). One of

the most striking aspects of this overall data pattern was
the fact that
these reliable differences in social problem solving emerge

after only the briefest,
casual exposure to cigarettes (NS versus ES groups).

Apparently the act of
smoking per se, irrespective of amount, duration or

frequency of cigarette use, is
associated with a subtle yet discernible shift in self and

situational perceptions.
This finding corroborates other research that documents the

rapid onset of
labeling of one's self as a smoker once cigarette use

begins (D'Onofrio, 1983a).
Equally interesting was the fact that self-designated

ex-smokers were also
reliably differentiated from continuing smokers on several

dimensions. It
appears that the initiation of smoking as well as its

cessation is related to a
powerful restructuring of one's perceptions, expectations

and actions in tempting
situations. The only apparent exceptions to this consistent

rank-ordered dif
ferentiation (e.g., NS > ES > XS > RS) were: (1) the

uniform importance
assigned by all groups to the perceived instrumental value

of cigarettes for
enhancing feelings of autonomy, maturity and affiliation

(needs issues); and
(2) the fact that the ES, XS, and RS groups, all of whom
had been involved to
some degree with cigarettes, were largely undifferentiated

in their tendency to
yield to their friends' wishes (peer strategy) even when

(especially for the ES and
XS groups) this meant rejecting parental advice. Causal

Dynamics Given the above evidence for consistent
differences in problem-solving
approaches among the various smoking status groups, we can

now consider the
potential causal role of these various factors in the

dynamics of smoking uptake
and resistance. As will be recalled, this inquiry was

pursued in the context of the
causal model presented in Figure 1. In essence, we asked:

How substantial is the
causal dependency of smoking intentions on problem-solving
factors, and is this
source of variance sufficiently robust to reduce the

initially dominant influence of
temptation on intentions? In order to determine the most

discriminating set of problem-solving factors,
all ten variables listed in Table I were entered into a

preliminary multiple regres
sion analysis using stepwise inclusion criteria with

intentions to smoke as the cri
terion variable. Six significant factors emerged. In

descending order of import
ance these were: affecting consequences (~ = -.272, R2 =

.166); needs con
sequences (~ = .136, R 2j = .031); interpersonal strategies

(~ = -.118, R 2j = .020);
self/peer strategies (~ = .143, R 2j = .019); health

issues (~ = -.109, R 2j = .011)
and decision-making strategies (~ = -.050, R 2j = .002).

As a group, these factors
accounted for some 24.92% of the variance in intentions to

smoke, with the
remaining four factors together contributing nothing

additional to the prediction
equation, F(4,3816) = .559 (p < .05). Next, these six

problem-solving variables
were entered as a block in a hierarchical regression

analysis following temptation
in the temporally ordered sequence portrayed in Figure I.

By employing a path
analytic interpretation, we can assess the direct effects

of both the temptationSalient Issues Anticipated
Consequences Prosocial Strategies
Figure 2 Path diagram of the moderating effects of
problem-solvjng skills on the relationship between
temptation and intentions to smoke. All pathways
significant, p < .05. Smoking Temptation Intention to
Smoke
intention and the problem-solving-retention linkages, as

well as the moderating
role of problem-solving elements in altering the

temptation-intention relationship
(e.g., temptation strategies intentions). The magnitude and

direction of the path
coefficients associated with each of the presumed causal

linkages are presented in
Figure 2. All pathways were significant at the a = .05

level. Inspection of the fITst column of Table 2 indicates
a significant zero-order
correlation between temptation and intentions to smoke of

.569 (p < .05) that
decomposes into a direct effect from temptation (p = .440,

p < .05), and into a
total indirect influence moderated through the

problem-solving elements of .129.
Among all the mediating problem-solving variables,

affective consequences was
the most substantial transmitter of temptation, accounting

for approximately half
(.058) of the total indirect influence on intentions.
Table 2
Decomposition of the Zero-Order Correlation of Temptation

Level With the Intention to
Smoke Effect % Effect Zero-order correlation (r) .569*

Direct (p)a .440* 77.3% Indirect via: Health issues
.019 3.3% Decision-making strategies .008 1.4%
Interpersonal strategies .014 2.5% Self/peer strategies
.019 3.3% Affective consequences .058 10.2% Needs
consequences .011 1.9% Total indirect .129 22.75 Total
causal .569 Noncausal
Note. a standardized regression coefficient; *p < .05.

Next, consider the direct influence of problem-solving
elements on intentions
to smoke. Figure 2 indicates that all six components made

significant contribu
tions to variations in intentions, irrespective of

temptation level, p < .05. Thus,
for instance, to the extent that health issues were salient

to an individual, inten
tions to smoke were correspondingly reduced (p = -.089, p

< .05). The same
inhibiting influence was seen in the presence of

decision-making strategies (p =
-.046), in the possession of interpersonal strategies (p =

-.049), and to the degree
that smoking represents a violation of affiliative bonds

(affective consequences)
(p = .180), p < .05. Contrariwise, intentions to smoke

increased to the extent
smoking was perceived as leading to need-fulfillment

regarding autonomy and
maturity (p = .048) and to the degree that subjects were

likely to defy authority
and comply with perceived peer group wishes (p = .096), p <

.05. The results of incremental F-tests (not tabled)
provided an overall summary
of this path analysis. As the first predictor variable

entered, temptation account
ed for some 79.8% (R 2j = .324) of the total explained

variation in intentions
(R 2j = .406) through all paths of influence, both direct

and indirect, F( I ,4090) =
2230.91, P < .01. The six combined problem-solving sources

added the
remaining 22.2% of the explained variance (R2j = .082),

F(6,4084) = 93.96, P <
.01. These results establish not only the importance of the

aggregated problem
solving variables as direct sources of influence on

intentions to smoke in the face
of temptation, but also indicates something of their

indirect role in mediating
temptation. This latter causal source amounted to

approximately one-fifth (R2j =
.059) of the explained variance attributable to temptation

(R 2j = .324). Finally, consider the evidence on whether
or not the elements of problem
solving under investigation contribute equally to smoking

intentions regardless of
the age and smoking history of the individual. As to age

level, pairwise compari
sons of the seven direct effect regression slopes across

the three age levels
revealed essentially identical findings, all t (00) <

1.96, P < .05. Thus, for exam
ple, while the affective consequences of a smoking

decision may be more salient
to younger children (see Table 1), the causal mechanisms

by which such varia
tions in expected parental, peer and self-upset translate

into intentions are no
more evident among 11to 12-year-olds among older

youngsters. Nor are the
above findings subject to qualification depending on the

individual's smoking
history. Pairwise contrasts between the regression slopes

of all four smoking
status categories for each direct linkage portrayed in
Figure 1 revealed no differ
ences in the causal role of any of the problem-solving

components in determining
smoking intentions, all t (00) < 1.96, p < .05. For

instance, while RS individuals
may have less well-developed interpersonal skills to

deflect smoking temptation
and express greater intentions to smoke (Table I), we

conclude from these insig
nificant comparisons of regression slope that such

deficiencies in problem
solving approaches exert no greater causal impact on

smoking intentions among
smokers than among adolescents with little or no smoking

experience. Several
educational implications emerge from these data. First,

according to the empiri
cally confirmed linkages portrayed in Figure 1, any

interventions that act to
increase the saliency of health issues and the

availability of skills in inter
personal, prosocial coping strategies can be expected to

decrease smoking inten
tionality, regardless of the temptation value of a given

circumstance. Likewise,
any instruction that decreases the tendency toward defiance

of authority and the
uncritical acceptance of cigarettes as evidence of

maturity will also dampen
intentions. Moreover, it is possible-although no empirical

demonstration is yet
available-that interventions might simultaneously act to

alter the magnitude of
causal relationships between the various problem-solving

elements and inten
tionality. Thus, for example, in theory, instruction might

act to increase the cau
sal importance of various problem-solving elements (e.g.,

interpersonal skills) for
reducing the likelihood of smoking, as well as to raise

the absolute level of skill
proficiency. In such an instance, instruction would be

doubly effective in
controlling intentions to smoke. Second, the indirect

linkages by which problem-solving components mediate
the influence of temptation on intentions (temptation --+

problem solv
ing --+ intentions) is another prime source for

educational interdiction. In effect,
by enhancing problem-solving skills, the instructional

planner can control a pro
portion of the variability in intentions that would

otherwise be subject totally to
the whims of temptation. However, in order to maximize this

potential source of
control, the inhibitory effects of temptation on the

various problem-solving com
ponents must be offset, as reflected in Figure 2

(temptation --+ problem solving).
For instance, consider the all-too-common human tendency to

discount rational,
self-interested concerns in the face of temptation and

pressure, a tendency which
often takes the form of denying and minimizing health

risks (Covington &
Omelich, in press). These dynamics are wen illustrated by
the behavior of the
subjects regarding the health issue variable. In effect,

health issues became less
salient to the decision-making process as self-perceived

temptation increased (p
= -.218, p < .05). These data underscore the need to teach

more than health facts
alone since such information is subject to avoidance,

minimization and distortion,
a process thought to be caused by strong emotional needs

including the denial of
illness and death (DeLong, 1970). The same distorting

influence of temptation
can also be seen at work in the suspension of otherwise

inhibitory decision
making propensities (p = -.182), interpersonal strategies

(p = .332), p < .05. Educationally speaking, one
reasonable intervention strategy is to enhance
the saliency of problem-solving considerations as a

direct, inhibitory influence on
intentions to smoke, and simultaneously to reduce the

distorting effects of temp
tation on rational, problem-solving considerations. At the

same time, those
mechanisms associated with temptation that lead to

increased perceptions of the
utility of smoking (need fulfillment; p = .229, p < .05)

and to a greater willing
ness to abide by perceived peer wishes (peer orientation;

p = .195, P < .05) must
also be made targets of special corrective treatment.

GENERAL DISCUSSION The overall results of this study offer
support for the view that a social
problem-solving orientation represents a potentially viable
approach to anti
smoking intervention. First, not only did various

problem-solving elements such
as perceived issues, prosocial strategies and expected

consequences influence
intentions to smoke in their own right, but they also

acted to offset the con
siderable impact of temptation as a primary cause of

smoking uptake. Second,
this problem-solving dynamic appeared to operate in

essentially the same fashion
for smokers and nonsmokers alike, a parallelism that

suggests the broad applica
bility of such an approach whether the educational goal is

that of smoking
prevention or cessation. A further implication to be

drawn from these data is that to be most effective,
strategies for intervention must not focus solely on

health issues, but must also
address the need of young people to satisfy naturally

occurring motives of
personal/social significance. If instruction remains solely

at the level of health
facts, then educators will be forced to rely heavily on the

child's will-power and
sometimes on poorly developed cognitive controls to

motivate a nonsmoking
decision. Although the health issues factor in this study

made a distinct impact
on smoking intentions, and while smokers appeared

disproportionately disposed
to reject health facts, health issues nonetheless were a

relatively minor causal
source when compared to other factors such as anticipated
affective conse
quences. This is not to suggest that health facts are

unimportant or irrelevant, but
only that they must be placed in proper perspective as only

one of several
promising foci for intervention. How, then, can the role

of problem solving be best conceptualized without
underemphasizing the role of health facts and in ways that

are suggestible of spe
cific types of intervention activities? An overall

organizing principle is strongly
suggested by the present data: informed self-interest.

Concern for the social and
personal consequences of smoking emerged as a major

predictor of intentionality
for both smokers and nonsmokers. Inspection of the items

that make up the af
fective consequences factor in the principal components

analysis used to derive
the scales in this study indicates that concern for one's

self-image has the highest
factor loading (.765), followed by concern for parental

reactions (.643) and peer
reactions (.597) to one's smoking. In this context,

self-interest involves the
potential violation of self-integrity by actions that are

inconsistent with one's
ideal self as well as the need to maintain family harmony

and yet to maximize
peer acceptance. These are the problem elements that must

be balanced if the
child is to avoid smoking. For nonsmokers these elements

are compatible with
the child's self-expectations, as well as with those of

his or her family and peers.
By comparison, most smokers suffer a sense of misalignment

regarding smoking
values. Their parents' advice is often at odds with

prevailing peer values and with
their own emerging perceptions of self. This conflict may

range in emotional tone
from feeling mildly frustrated to an expression of outright

antagonism toward
parents, if we are to judge from the data. The behavior of

the peer strategy factor
underscores this point. Not only do many smokers discount

parental values, but
for some this amounts to an outright rejection in favor of

anarchistic freedom
(e.g., "I'll do what I like, after all it's my life").

Such a conflict makes it all the
more difficult to disentangle health issues from their

emotional overlay. When
intrafamily stress is combined with a greater acceptance

of smoking among one's
peers, and with compelling, if false, beliefs about the

instrumentality of smoking,
then the scales are tipped decidedly in favor of a decision

to smoke. This analysis suggests that successful
anti-smoking intervention should
involve at least three kinds of problem-solving training.

First, it is important to
enhance the capacity of youngsters to identify the

essential issues of self-interest
that arise in any smoking-related decision including: (a)

self-consistency and
valued self-image; (b) the desirability of harmonious
relations with parents;
(c) the need to maintain positive peer relationships; and

(d) the necessity of
continued health maintenance. A second intervention focus

strongly implied by the significant interpersonal
strategies factor in this study is to provide youngsters

with repeated practice in
generating resolutions to tempting situations that balance

each of these self
interest elements. Or, as young students best understand

the task, "What can I do
to keep my friends, feel good about myself, get along with

my parents-all with
out smoking?" The item content of the interpersonal

strategy factor suggests
some of the specific skills that such dilemma resolutions

might embody: "Do
something so that others do not feel they have to smoke";

"Talk about my
decision so that everyone will still be friends no matter

what the others do"; and
"Try to talk the others out of smoking." In this

connection one fruitful direction
for effective decision-resolution training might involve

practice in establishing a
cost/benefits hierarchy; in effect, helping children

determine which resolutions
are better or worse given all the likely consequences of

each. A third problem-solving focus implied by the present
study is to dismantle
health rationalizations that tend to discount the risk of

smoking especially when
temptation is great. Some evidence (Covington, 1981)
suggests that youngsters
often smoke without explicit intentions to do so and

without any particular
thought given to it (Le., simply because the cigarettes are

available). In cases like
this, when there is no intention to smoke it seems

unreasonable from the ado
lescent point of view that they should suffer negative

health consequences. The
responsibility for smoking behavior is either attributed to

others (e.g., "the group
made me do it") or is compartmentalized away from

intentions (Covington &
Omelich, in press). Needless to say, dealing with this form

of denial is a complex
challenge, especially among children whose sense of future

is typically compres
sed and may be dominated by prelogical thought patterns

that stress magical,
unrealistic thinking. In this connection, one promising

instructional approach is
the use of probability games designed to simulate those

real-life factors that
affect the youngsters' (players') acceptance or rejection

of smoking risks. Such
simulations might involve a sequential series of life

transitions in which the deci
sion to smoke in an early round of play (in order to gain

peer-group acceptance in
the teenage years) puts the player at a handicap in future

rounds (as an adult)
when the objective is now to establish financial

independence, a task made un
certain by increased health risks, ineligibility for
certain jobs, and unexpected
medical expenses. Here delayed cause-and-effect dynamics

can be discovered
and the implications understood by the early adolescent,

all without preaching
Ajzen, I., & Madden, 1. T. (1986). Prediction of


Experimental Social


promoting behavior. In 1. Madden, S. Matthysse, & 1.

Barchas (Eds.), Adaptation,
learning and affect. New York: Raven.
Best, J. A., & Hakstian, A. R. (1978). A

situation-specific model for smoking behavior.
Addictive Behaviors, 3, 79-92.
Botvin, G. 1. (1983). Life skills training: A

self-improvement approach to substance
abuse prevention. Teacher's Manual. New York: Smithfield.
Botvin, G. J., & Eng, A. (1980). A comprehensive

school-based smoking prevention
program. Journal of School Health, 50, 209-213.
Botvin, G. 1., & Eng, A. (1982). The efficacy of a

multicomponent approach to the
prevention of cigarette smoking. Preventive Medicine, 11,

199-211.
Botvin, G. J., Eng, A., & Williams, C. L. (1980).

Preventing the onset of cigarette
smoke through life skills training. Preventive Medicine, 9,

135-143.
Cohen, S., Lichtenstein, E., Mermelstein, R., Kingsolver,

K., Bauer, 1. S., & Kamarck,
T. W. (1988). Social support interventions for smoking

cessation. In B. H. Gottlieb
(Ed.), Marshaling social support. Formats, processes, and

effects (pp. 211-240).
Beverly Hills, CA: Sage.
Covington, M. V. (1981). Strategies for smoking prevention

and resistance among
young adolescents. Journal of Early Adolescence, I,

349-356.
Covington, M. V. (1986). Instruction in problem-solving

planning. In S. L. Friedman, E.
K. Scholnick, & R. R. Cocking (Eds.), Blueprints for

thinking: The role of planning
in cognitive development, (pp. 469-511). New York:

Covington, M. V., D'Onofrio, C. N., Thier, H. D., Schnur,

A. E., & Omelich, C. L.
(1983). Risk and youth: Smoking (RAY:S) prevention program.

Lawrence Hall of
Science, University of California at Berkeley.
Covington, M. V., & Omelich, C. L. (1986). Family vs. peer

influences on adolescent
smoking decisions: A developmental approach. Unpublished

manuscript, Department
of Psychology, University of California at Berkeley.

Covington, M. V., & Omelich, C. L. (1988). I can resist
anything but temptation:
Adolescent expectations for smoking cigarettes. Journal of

Applied Social
Covington, M. V., & Omelich, C. L. (in press). The

perceived costs and benefits of cig
arette smoking among adolescents: Need instrumentality,

self-anger and anxiety fac
tors. In D. G. Forgays, T. Sosnowski, & K. Wrzeniewski

(Eds.), Anxiety: Recent
developments in self-appraisal, psychophysiological and

health research. New York:
Hemisphere.
Delong, W. R. (1970). Individual differences in patterns of

anxiety arousal, stress-rele
vant information and recovery from surgery. Unpublished

doctoral dissertation,
University of California at Los Angeles.


behavioral intentions. Health
Education Research, 3,273-282.
D'Onofrio, C. N. (1983a, April). Smoking among children:

The nature of the problem.
Paper presented at the Annual Convention of the Western

Psychological Association,
San Francisco.
D'Onofrio. C. N. (1983b). Review of evaluated school-based

anti-smoking programs.
Unpublished manuscript, University of California at

Berkeley.
D'Onofrio, C. N., Thier, H. D., Schnur, A. E., Buchanan, D.

R., & Omelich, C. L.
(1982). The dynamics of adolescent smoking behavior. World

Smoking and Health,
7, 18-24.
Evans, R. I., Hansen, W., & Mittlemark, M. B. (1977).

Increasing the validity of self
reports of behavior in a smoking-in-children

investigation. Journal of Applied
Evans, R. I., Henderson, A., Hill, P., & Raines, B.

(1979). Smoking in children and
adolescents: Psychosocial determinants and prevention

strategies. In N. A.
Krasnegor (Ed.), The behavioral aspects of smoking.

Washington. DC: NIDA
Research Monograph 26.
Fishbein, M. (1977). Consumer beliefs and behavior with

respect to cigarette smoking:
A critical analysis of the public literature. A report

prepared for the staff of the
Federal Trade Commission.
Garcia, M. E., Schmitz, 1. M., & Doerfler, L. A. (1990). A

fine-grained analysis of the
role of self-efficacy in self-initiated attempts to quit

smoking. Journal of Consulting
and Clinical Psychology, 58(3),317-322.
Gilchrist, L. D., Schinke, S. P., & Blythe, B. 1. (1979).

Primary prevention services for
children and youth. Children and Youth Services Review, J,

379-391.
Herrnstein, R. J., Nickerson. R. S., Sanchez, M .• & Swets,
1. A. (1986). Teaching
thinking skills. American Psychologist, 4 J, 1279-1289.
Jones, 1., Piper, G. W., & Matthews, V. L. (1970). A

student-directed program in
smoking education. Canadian Journal of Public Health, 6 J,

253-256.
Kim, S. (1982). Feeder area approach: An impact evaluation

of a prevention project on
student drug abuse. International Journal of the

Addictions, J 7, 305-313.
Kirk, R. E. (1968). Experimental design: Procedures for

the behavioral sciences.
Belmont, CA: Brooks/Cole.
Leventhal, H., & Cleary, P. D. (1980). The smoking

problem: A review of the research
and theory in behavioral risk modification. Psychological

Review, 88. 370-405.
Nickerson, P. S .• Perkins, D. N., & Smith, E. E. (1985).

The teaching of thinking.
Pedhazur, E. J. (1982). Multiple regression in behavioral

research: Explanation and
prediction (2nd ed.). New York: Holt, Rinehart & Winston.
Piper, G. W., Jones, 1. A .• & Matthews, V. L. (1974). The

Saskatoon smoking project
The model. Canadian Journal of Public Health, 65, 127-129.
Schinke, S. P., Gilchrist, L. E., Snow, W. H., &

Schilling, R. 1. (1985). Skills building
methods to prevent smoking by adolescents. Journal of

Adolescent Health Care,
6(6), 439-444.
Sutton, S. (1989). Smoking attitudes and behavior:

Applications of Fishbein and
Ajzen's theory of reasoned action to predicting and

understanding smoking decisions.
In T. Ney & A. Gale (Eds.), Smoking and human behavior

(pp. 289-312). Chichester:
Wiley.
Williams, C. L., & Arnold, C. G. (1980). Teaching children

self-care for chronic
disease prevention: Obesity reduction and smoking

prevention. Parent Counseling
and Health Education, 8, 92-98.
Williams, C. L., Eng, A., Botvin, G. J., Hill, P., &

Wynder, E. (1979). Validation of
students' self-reported cigarette smoking status with

plasma cotinine levels.
American Journal of Public Health, 69, 1272-1274. This

page intentionally left blank
V Self-Efficacy, Physical Symptoms, and
Rehabilitation of Chronic Disease

Prentice-Hall.
Bandura, A. (1990). Perceived self-efficacy in the

exercise of personal agency.
Applied Sport Psychology, 2, 128-163.
Bandura, A., & Cervone, D. (1983). Self-evaluative and

self-efficacy mechanisms gov
erning the motivational effects of goal systems. Journal

Psychology,45,1017-1028.
Bersheid, E. (1983). Emotion. In H. Kelley, E. Bersheid, A.

Christensen, J. H. Harvey,
T. L. Huston, O. Levinger, E. McClintock, L. A. Peplau, &

D. R. Peterson (Eds.),
Close relationships (pp. 110-168). New York: Freeman.
Boutcher, S. H., & Landers, D. M. (1988). The effects of

vigorous exercise on anxiety,
heart rate, and alpha activity of runners and nonrunners.

Psychophysiology, 25, 696
702.
Bower, O. H. (1981). Mood and memory. American

Brown, J. D. (1991). Staying fit and staying well:

Physical fitness as a moderator of life
stress. Journal of Personality and Social Psychology, 60,

555-561.
Cioffi, D. (1991). Beyond attentional strategies: A

cognitive-perceptual model of
somatic interpretation. Psychological Bulletin, 109,25-41.

Coyne, J. c., & Smith, D. A. F. (1991). Couples coping with
a myocardial infarction: A
contextual perspective on wives' distress. Journal of

Croog, S., & Levine, S. (1982). Life after a heart attack:

Social and psychological
factors. New York: Human Science Press.
DeBusk, R. F., Blomqvist, C. G., Kouchoukous. N. T .•

Luepker. R. V .• Miller. H. S.,
Moss, A. J .• Pollock, M. L., Reeves. T. 1.. Selvester. R.

H., Stason, W. B .• Wagner. G.
S., & Willman. V. L. (1986). Identification and treatment

of low-risk patients after
acute myocardial infarction and coronary-artery bypass

graft surgery. New England
Journal of Medicine. 314. 161-166.
Emmons. R. A. (1986). Personal strivings: An approach to

personality and subjective
well-being. Journal of Personality and Social Psychology,

51. 1058-1068.
Ewart. c.K. (1989). Psychological effects of resistive

weight training: Implications for
cardiac patients. Medicine and Science in Sports and

Exercise. 21. 683-688.
Ewart. C. K. (1990). A social problem-solving approach to

behavior change in coro
nary heart disease. In S. Shumaker, E. Schron, & J. Ockene

(Eds.), Handbook of
health behavior change (pp. 153-190). New York: Springer.
Ewart, C. K. (199 I). Social action theory for a public

health psychology. American
Psychologist. 46.931-946.
Ewart, C. K.. Stewart. K. J., Gillilan. R. E., & Kelemen.

M. H. (1986). Self-efficacy
mediates strength gains during circuit weight training in

men with coronary artery
disease. Medicine and Science in Sports and Exercise.

18.531-540.
Ewart, C. K.. Stewart, K. J., Gillilan. R. E., Kelemen. M.

H .• Valenti. S. A .• Manley.
J. D., & Kelemen, M. D. (1986). Usefulness of

self-efficacy in predicting
overexertion during programmed exercise in coronary artery

disease. American
Journal of Cardiology. 57. 557-561.
Ewart, C. K.. & Taylor. C. B. (1985). The effects of early

postmyocardial infarction
exercise testing on subsequent quality of life. Quality of

Life and Cardiovascular
Care, I. 162-175.
Ewart. C. K .. Taylor. C. B., Reese, L. B., & DeBusk, R. F.

(1983). Effects of early
postmyocardial infarction exercise testing on

self-perception and subsequent physical
activity. American Journal of Cardiology, 51. 10761080.
Folkins. C. H., & Sime, W. E. (1981). Physical fitness

training and mental health.
American Psychologist, 36,373-389.
Frijda, N. H., Kuipers. P .• & ter Schure, E. (1989).

Relations among emotion. appraisal,
and emotional action readiness. Journal of Personality and

Social Psychology, 57,
212-228.
Kavanagh. D. 1.. & Bower, G. H. (1985). Mood and

self-efficacy: Impact of joy and
sadness on perceived capabilities. Cognitive Therapy and

Research. 9, 507-525.
Lazarus, R. S. (1991). Progress on a

cognitive-motivational-relational theory of emo
tion. American Psychologist, 46, 819-834.
Martin, 1. E., Dubbert, P. M., Katel1, A. D., Thompson. 1.

K .. Raczynski, J. R., Lake.
M., Smith, P.O., Webster, J. S., Sikora, T., & Cohen, R.

E. (1984). Behavioral con
trol of exercise in sedentary adults: Studies I through 6.

Clinical Psychology, 52. 795-811.
Miller, N. H., Haskell, W. L., Berra. K., & DeBusk, R. F.

(1984). Home versus group
exercise training for increasing functional capacity after

myocardial infarction.
Circulation, 70, 645-649.
O'Connor, G. T., Buring, J. E., Yusuf, S .. Goldhaber, S.

Z., Olmstead. E. M.,
Paffenbarger, R. S., & Hennekens. C. H. (1989). An

overview of randomized trials
of rehabilitation with exercise after myocardial

infarction. Circulation. 80, 234-244.
Pennebaker, J. W. (1982). The psychology of physical

symptoms. New York: Springer.
Ross, C. E., & Hayes, D. (1988). Exercise and psychologic

well-being in the commun
ity. American Journal of Epidemiology, 127,762-761.
Steptoe, A., & Cox, S. (1988). Acute effects of aerobic

exercise on mood. Health


husband's capability soon after
clinically uncomplicated myocardial infarction. American

Journal of Cardiology, 55,
636-628.
Taylor, C. B., Sallis, J. F., & Needle, R. (1985). The

relation of physical activity and
exercise to mental health. Public Health Reports, /00,

195202.
Tellegen, A. (1985). Structures of mood and personality

and their relevance to assessing
anxiety, with an emphasis on self-report. In A. H. Tuma &

1. D. Maser (Eds.),
Anxiety and the anxiety disorders (pp. 681-706). Hillsdale,

NJ: Erlbaum.
Watson, D. W., Clark, L. A., & TeUegen, A. (1988).

Development and validation of
brief measures of positive and negative affect: The PANAS

scales. Journal of
Watson, D. W., & Pennebaker, J. W. (1989). Health

complaints, stress, and distress:
Exploring the central role of negative affectivity.

Psychological Review, 96, 234-254. Author Notes
Preparation of this manuscript was supported in part by
grants ROl-HL36298
and ROI-HL45139 from the National Heart, Lung, and Blood

Institute of the
National Institutes of Health, Bethesda, Maryland.

PERCEIVED SELF-EFFICACY IN SELF-MANAGEMENT OF CHRONIC
DISEASE Halsted Holman and Kate Lorig Chronic disease,
now the most prevalent form of disease in the United
States, differs from acute disease in many ways. One of
the most important is the potential for self-management by
patients. Appropriate self-management is based upon a
partnership between the patient and health professionals
in which each takes responsibility for portions of the
management. For patients, this requires learning new skills
and assuming new responsibilities. Growing evidence
indicates that perceived self-efficacy to cope with the
consequences of chronic disease is an essential
contributor to developing self-management capabilities,
and that perceived self-efficacy can be rapidly enhanced by
appropriate learning experiences. Thus perceived
self-efficacy is an important personal attribute in the
maintenance of health. Enhancing perceived self-efficacy
should be an important ingredient of the provision of
health care.
The central thesis of this chapter holds that, in the

presence of a chronic disease,
many types of self-management practices are both feasible

and beneficial, and
that perceived self-efficacy to execute those practices and

to manage the conse
quences of the disease improves the outcomes greatly. In

order to explain and
substantiate the thesis, we must make clear each of the

component concepts.
Therefore this chapter is organized into sections which

define and discuss (I) the
nature of chronic disease and its management (2)

characteristics of self
management and (3) the interactions among self-efficacy,

self-management and
that the reader can explore the subject more fully.

However, this chapter is not a
comprehensive review. It presents conceptual approaches,

citing supporting
empirical evidence. The Nature of Chronic Disease and its
Management
Chronic disease has become the most prevalent form of

disease in the United
States (Rice & Feldman, 1983; Verbrugge, 1984). It is the

principal source of
disability and a major cause of escalating health care

expenditures (Colvez &
Blanchet, 1981). Because the prevalence of chronic dise!lse

increases with age,
the aging of the population has contributed to increases

in prevalence. However,
in recent years, the prevalence of chronic disease has

increased for virtually every
age group in the population (Rice & LaPlante, 1988). While

chronic disease
typically arises spontaneously, it can also be created by

contemporary treatment
of acute illness. For example, modem treatment of an acute

heart attack which
previously would be fatal can now result in an individual

with chronic heart
disease. Usually, chronic disease cannot be cured. It must

be managed over
time. Treatment may correct certain biological

abnormalities, can ameliorate
some consequences of chronic disease, and may prevent

overall deterioration, but
it does not terminate the disease. Commonly it is not

possible to predict
accurately what effect a particular treatment or management

technique will have.
The effect may be beneficial, neutral or even harmful;

usually only a trial of
treatment will reveal the effects. The intensity of a
chronic disease typically fluctuates of its own accord.
That
is, the disease process is not relentless but rather has

periods of greater and lesser
severity. Treatment influences the fluctuations, usually

for the better. However,
if a treatment has an adverse effect, an additional

medical problem is actually
created by the management effort. The chronicity of a

disease creates problems beyond the specific consequenc
es of the particular biological abnormality (e.g., the

symptoms such as thirst and
urination due to high blood sugar in diabetes or pain and

reduced exercise toler
ance due to reduced blood supply from hardening of the

arteries in arteriosclero
sis). The additional consequences are also a result of the

disease and its treat
ment, and fall in many categories: fatigue, depression,

unfavorable employment
and financial circumstances, reduced social activities,

family conflict. At times,
these additional consequences react back on the disease

process to make matters
worse. Examples would be the emotional refusal of a

diabetic to secure an
appropriate diet with resultant worsening of the diabetes,

or the financial inability
of a physically handicapped person to obtain appropriate

physical rehabilitation
with resultant worsening of the handicap. It is common to

refer to the biological
abnormalities of a disease as the disease process and the
composite of that di
sease process and all of its consequences as an illness.

Thus a chronic illness has
many components and these components can interact with one

another to worsen
or lessen the total illness effect. Many illness

attributes are common across chronic diseases. That is,
while
the underlying biological abnormalities of chronic

diseases may be quite distinct,
the illness components such as those mentioned above, and

their potential inter
actions, have many similarities. Therefore, it is

appropriate to examine chronic
diseases as a group, seeking common themes in both their

manifestations and
their management. A useful first step in understanding

chronic diseases is recognition that they
are different from acute diseases. Table 1 enumerates some

of the distinctions.
For example, acute diseases usually result from

identifiable, abrupt and
potentially reversible specific events (e.g., infection,

injury, blood vessel rupture
or occlusion). By contrast, initiating events for most

chronic diseases are
obscure. The disease processes appear to emerge over time

from protracted inter
action among environmental, genetic and behavioral risk

factors. Once
established, chronic disease may be further altered by

interactions among the
disease process and its consequences (see above).

Table I
Disease Characteristics
Onset
Duration
Cause
Diagnosis and
prognosis
Technological
intervention
Outcome
Uncertainty
Knowledge Acute Disease Abrupt Limited Usually single

cause Commonly accurate Usually effective (laboratory
testing, imaging, medication, surgery) Cure likely with
return to normal health Minimal Profession
knowledgeable; laity inexperienced Chronic Disease/
Illness Commonly gradual Lengthy; indefinite
Multivariate causation of both disease and illness,
changing over time Diagnosis often uncertain; prognosis
obscure Commonly indecisive; adverse effects frequent No
cure; management over time necessary Pervasive
Profession and laity partially and reciprocally
knowledgeable For purposes of this chapter, the most
relevant difference between acute and
chronic disease lies in the role of the patient and

family, particularly in relation to
management. For acute diseases, the patients are usually

inexperienced and are
dependent upon health professionals for knowledge,

decisions and therapy. With
chronic illness, however, the patient commonly becomes the

most knowledgeable
person concerning both consequences of disease and effects

of therapy. At times,
the physician is dependent upon the patient even for

decisions in changing the
medical management. For instance, patients are often best

equipped to sense the
effect of changes in medication dose, to perceive the

relevance of a particular
management maneuver to the state of the illness, and to

decide to persevere in the
face of sub-optimal treatment outcomes. Appropriate

management of chronic di
sease requires development of a partnership between the

patient and the
physician. Patients, families and physicians come to the

relationship created by chronic
disease with widely different backgrounds, levels and types

of understanding,
and suppositions. Patients initially have almost no

experience with chronic di
sease, are beset by fears, and conceptualize the illness

through a variety of cul
tural and metaphorical understandings. Within these

understandings, symptoms
may have different individual and symbolic meanings. For

example, pain may be
viewed as a routine manifestation of disease, or as a

symbol of worsening and
reason for pessimism, or even as a punishment.

Understandings of the symbo
lisms create explanatory models for the patient about his

or her circumstances
and fate. Thus patients with the same diseases may have
strikingly different ill
ness patterns and explanatory models (Lewis & Daltroy,
1990). Further, patients
commonly seek certainty in explanation, prognoses and

treatment. Physicians,
on the other hand, generally share a biologically-based

explanatory model and
related treatment methods. While familiar with wide

variation in illness patterns
and responses to therapy, and hence with uncertainty, they

tend to disregard that
which is outside their explanatory model. The views of the

patient and the physician affect the mood of each, the
willingness of each to act, and the types of actions

taken. Effective management
of chronic disease requires building a partnership based on

merged understand
ings and actions, a process which requires inquiry,

interpretation, learning and
negotiation. Gradually, as experience with the chronic

disease grows, patients' under
standing of their particular circumstances grows. This

growth of understanding
includes not only the physical effects of the disease and

treatment but also the
consequences in all of the affected aspects of living.

Further, the patient and
family learn what can be done to ameliorate undesirable

consequences through
such steps as modification of life styles and living

routines, use of rehabilitation
and assistive devices, and drawing upon community

resources. Characteristics of Self-Management
Historically, in the health field, there have been two
general lineages from
which contemporary self-management arose: medical therapy
and public health
practices. Long before the days of medical

professionalism, individuals and
families engaged in various forms of treatment and/or

management of their health
problems (Starr, 1982). Similarly, there is a long

tradition of community action
to protect public health such as improving water and food

supplies, providing
appropriate housing, and specifying behavior such as

quarantine and immuniza
tion to prevent spread of disease. In the present era of

predominant chronic
disease, these traditions tend to modify and meld,

creating new roles for both
patients and health professionals. The new roles are

exemplified by the goal of
partnership. In our view, the patient's role is to engage

in the maximum feasible
self-management; the health professional's role is to

develop and facilitate that
self-management, teaching and providing expert knowledge.

Self-management means having, or being able to obtain, the
skills and
resources necessary to best accommodate to the chronic

disease and its conse
quences (Holroyd & Creer, 1986). The skills and resources

are both general
across chronic diseases and specific to particular

diseases. Because chronic
diseases and illnesses are unpredictable and may fluctuate

even on a day-to-day
basis, they often must be "managed" on a daily basis.

Appropriate management
arises in the collaborative partnership of patients with

health professionals, and
collaboration distinguishes self-management from self-help

and twelve-step pro
grams which are fundamentally conducted by patients without

the participation
of health professionals. Understanding self-management of

chronic diseases requires knowledge of
the chronic disease, the relevant self-management practices

and the relationship
between self-management and professional medical

management. The scope of
this knowledge has been only partially sketched by the

foregoing discussion. It is
useful to look more closely at characteristics of chronic

diseases in order to
specify the role of self-management.
Identifying the Illness Course and Management Effects

Chronic diseases and their resulting chronic illnesses
typically fluctuate in
intensity over time, both spontaneously and in response to

therapies. Because
there is no cure, these fluctuations continue

indefinitely. Appropriate manage
ment requires appropriate responses to these fluctuations:

Increased disease
intensity usually requires increased treatment intensity

while the opposite is true
when the disease becomes less active. Figure 1 depicts this

situation in a hypo
thetical course of fluctuating disease intensity. Points

A, Band C represent the
same disease intensity. However, Point A reflects a time
when the disease trend
is worsening and the appropriate response would usually be

to intensify or
change therapy. Point B represents an improving trend in

the disease with a like
ly lesser need for therapy. Point C represents a disease

activity plateau in which
the minimum therapy necessary to maintain the stable

situation would probably
be appropriate. The central requirement in interpreting

such situations is accurate
identification of the disease trend as reflected by the

arrows. But knowing the trend alone is not sufficient.
The tempo or speed along the
trend is also essential. Is the disease worsening rapidly

or slowly? How quickly
must one react? How rapidly must the treatment take effect?
In Figure 1, the
tempo is depicted by the slope or steepness of the curve.

In both medical emer
gencies and chronic disease, trend and tempo are essential

in assessing a patient's
state; in emergencies they are compressed into minutes or

hours while in chronic
disease they play out over days, weeks and months.

However, the situation is not as simple as depicted in
Figure 1. Because the
disease process and the various components of the resulting

illness all fluctuate
over time, Figure 2 is a truer representation of the

clinical setting in which signs
and symptoms oscillate appreciably around the true course

of the disease and/or
illness. Sometimes, those oscillations may be so extreme
as to suggest a trend Severity
Figure I Severity Tim.
Figure 2
opposite to the true one. Also, the trends of the disease

and the illness may not
move in parallel; that is, a disease may be improving or

stable but a particular ill
ness component may worsen. An example could be pain in

chronic arthritis. The
intensity of pain, a major consequence of arthritis, can

vary with a person's emo
tional state; pain can worsen with emotional distress and

lessen when a person is
emotionally tranquil. Thus, though the arthritis may be

stable, external emotional
circumstances can aggravate a primary symptom of

arthritis. If the proper cause
of the increased pain is not identified and addressed,

treatment could be wrong.
The result could be lack of benefit and adverse treatment

effects. Thus it is not
only essential that trends in disease and illness be

identified but that oscillations
around the true trend be accurately interpreted. This

usualJy requires, in addition
to diagnostic perceptiveness, observation over time in

order to understand the
meaning of trends in sign and symptoms. Sometimes the

interpretation is assist
ed by diagnostic tests but commonly with chronic diseases

and illness, appro
priate use of time is all that is necessary to interpret

physical signs and symptoms
and thus to identify the trend and tempo. Indeed, when used
wisely, time can be
the most accurate, most available and least expensive tool

for the management of
chronic disease. A B C Time The discussion of Figures 1

and 2 reveals two of the most crucial attributes
of good management of chronic disease and illness, namely,

the identification of
trends and tempo, and the wise use of time to accomplish

that identification.
These two attributes of sound management provide the

conceptual base for self
management of chronic disease. While the physician and

other health profes
sionals are most knowledgeable about the biology of the

disease and the principal
treatment methods, the patient and the family are usually

in the best position to
identify the impact of the disease and treatment on the

individual, and to interpret
changes in the impact. Particularly if the patient develops

understanding of
(1) the nature of the disease, (2) the ways to interpret

signs and symptoms, (3) the
types of treatments and management procedures which are

available, and also
learns (4) to execute those aspects of management which

can be done personally,
a true partnership can emerge.
Basic Self-Management Skills In addition to learning about

the nature of chronic disease and illness, and
about medical management, there are seven basic areas of

skill which are central
to those aspects of management which can be done

personally. These are
(1) minimizing or overcoming physical debility, (2)

establishing realistic expec
tations and emotional responses to the vicissitudes of the

illness, (3) interpreting
and managing symptoms (4) learning how to judge the

effects of medications and
manage their use (5) becoming adept at ways to solve

problems as they arise,
(6) communicating with health professionals and (7) using

community resources
to advantage. Each will be discussed in sequence. 1. In

the presence of chronic disease, physical deconditioning
can result from
many causes: the disease itself (as when muscles or joints

are affected), reduced
activity (as when bedridden, handicapped or short of

breath), treatment (as when
using drugs which affect muscle strength like cortisone),

and poor nutrition.
Furthermore, the deconditioning creates its own symptoms

in that, when used,
deconditioned muscles, joints and tendons can cause pain

and accentuate fatigue.
The latter consequences are similar to those experienced by

a deconditioned
athlete who resumes training. Maintaining the maximum

possible physical fitness
is therefore a cardinal goal in management of any chronic

disease. Most people
with chronic diseases are benefitted by exercise and most

can establish exercise
programs that are within their tolerance and which lead to
maintaining or improv
ing physical capability. Such programs are often aided by

simple pain reducing
mechanisms such as the use of heat, mild analgesics like

aspirin and cognitive
pain control methods. Physical activity and stamina are

also aided by assuring
appropriate nutrition and sleep, both of which can be

achieved by various physi
cal, cognitive and medical strategies. 2. Achieving

appropriate expectations and new emotional adjustments can
be
the most important changes to be made by a patient with

chronic disease.
Patients with chronic illness often find themselves unable

to relieve persistent
symptoms and unable to participate in desired activities.

Simultaneously they are
often faced with uncertainty in diagnosis and prognosis.

For many, this uncer
tainty extends, through the fluctuations in illness

intensity, to their lives from day
to day. Planning becomes difficult; loss of work, of

social activities and of social
support systems may occur. In such a setting, patients may

develop a sense of
helplessness and depression. Health professionals can help

in various ways but much of an appropriate
response lies within the realm of self-management. It is

particularly helpful for
patients to learn what is known and is not known about

both the disease and its
therapy. This allows the patient to place himselflherself
in the context of prevail
ing knowledge. Understanding that context aids in

accommodating to uncertain
ty in a variety of ways: knowing that appropriate steps

are being taken; identify
ing other patients with similar problems whose experiences

can be helpful; learn
ing particular activities which prevent or compensate for

consequences of the
disease; learning how to confront most effectively a new

problem from the illness
or its therapy. Activities in each of these categories can

be very specific, for
example, minimizing pain by cognitive techniques,

maximizing mobility by use
of appropriate assist devices, reducing fatigue by both

exercise and rectifying
sleep disturbances, building new friendships with persons

with similar difficul
ties. Such approaches can be generally summarized under the

headings of learn
ing to function effectively and pleasurefully within the

limits imposed by the
disease while simultaneously taking actions to expand

those limits. As a
patient's repertoire of skills grows the distress of

uncertainty declines because of
a rising capacity to confront the variations which the

disease creates. 3. The most obvious effect of a chronic
disease is a symptom-pain, fatigue,
breathlessness, nausea. The disease may directly cause the

symptom, or the
symptom may be caused by a behavior such as exertion or by

an emotional upset,
or the symptom might arise from an error in medication dose

or from an adverse
effect of a drug. With chronic disease, patients often

become quite adept at inter
preting their symptoms. This skill is essential to

identifying appropriate action
and avoiding useless or harmful treatment changes. Often,

when there is uncer
tainty about symptom interpretation, a brief observation

period or a slight treat
ment change followed by observation will provide the

answer. 4. Most management of chronic disease includes the
use of medications. The
medications are not intended as a cure but rather as a

means of diminishing the
intensity of the disease or alleviating some of its

consequences. It is almost
always the responsibility of the patient or the family to

administer the treatment
on a day to day basis. Skills involved in medication

self-management include
those which assist in complying with the regimen and in

interpreting the effects
of the medication, in particular recognizing adverse

effects. When a new or
unanticipated event arises, the patient must be prepared to

describe and interpret
the event, to seek medical advice, and potentially to

alter the treatment program.
Patients can readily develop these skills but, to do so,

they must learn about
medications and their effects, and monitor their own

experiences carefully. 5. Problem solving is a major
self-management skill. One cannot predict in
advance the problems which an individual patient will face.

These can range
from the discomforts and disabilities of the illness to

inability to pay for health
care, to the need to change a residence, or to

difficulties in a marriage. Success in
dealing with new adverse situations greatly influences how

a patient accommo
dates to and manages an illness. Problem solving with

chronic disease is similar to that of virtually any other
situation. It includes defining the problem, identifying

alternative solutions,
choosing a solution to be tried, monitoring the results of

the trial, and testing
alternative solutions if desirable. While it is impossible

to teach patients all
potential solutions, it is possible to teach them the

generic skills for finding those
solutions. This involves teaching general problem solving

skills rather than solv
ing of specific problems. Though initially inexperienced

at resolving medical
problems, patients commonly become expert in knowledge of

their disease gener
ally. Building this knowledge into an understanding of how

medical problems are
solved, and providing access to appropriate consultation or

supervision, allows
the patient's common sense skills in problem solving to be

applied in a new area. 6. Health professionals are the
most common consultants used by people with
chronic illness. Unfortunately, all too often, care for

chronic disease is provided
like care for acute disease: Health professionals solve the
problems for a passive
patient; patients are neither encouraged nor assisted to

solve problems independ
ently. For the partnership necessary for the best possible

outcome from chronic
disease, patients need at least two essential skills.

First, the patient must know
when to seek advice, how to describe symptoms and how to

identify the trends
and tempo of the disease. Second, patients must be able

and willing to express
concerns and to negotiate with health professionals

concerning the next action.
Often this means questioning the physician, and accepting

or rejecting treatment
options. For their part, physicians must encourage

development of these skills by
the patient and must be willing to enter into a

partnership. When both parties
develop these skills, a process emerges which is both

efficient and satisfying. 7. Identification and use of
resources beyond those which are personal or are
provided by health professionals is the final general

self-management skill. Prob
lems caused by chronic illness are often soluble only

outside the health care sys
tem. These include, but are not limited to, use of exercise
facilities and programs
and use of special educational opportunities such as

libraries, community colleg
es and special education programs. Community groups may be

invaluable such as
senior citizen centers and organizations for people with
particular diseases. At
times, Meals-on-Wheels and Friendly Visitors can be

particularly helpful.
Prevention Prevention of disease is commonly perceived in

terms of direct prevention
such as sanitation, immunization, avoidance of substance

abuse and wearing
seatbelts. This is called primary prevention. There are two

other types of pre
vention, secondary and tertiary. Secondary prevention

involves avoiding a
disease once a risk factor is present. Examples include

modifying a diet when a
blood cholesterol level is high or reducing weight and salt

intake when blood
pressure is elevated. Tertiary prevention involves

preventing loss of independent
function once a disease is present. Examples include

maintaining general physi
cal strength and stamina in the presence of disease, use of

specific breathing pat
terns to improve oxygenation in lung disease and

elimination of household
obstacles over which one might fall when physically

handicapped. In the presence of chronic disease, secondary
and tertiary prevention are par
ticularly important. Some of the appropriate practices can

be learned from health
professionals while others entail the use of common sense

in confronting the
issues which arise in daily life. Prevention practices

designed to minimize the
consequences of chronic disease and maintain independent

living illustrate the
convergence of classical medical and public health

principles to create the most
effective management program for the patient and the

family. The afore-mentioned components of
self-management-understanding and
interpreting disease and illness patterns, using specific

self-management practices
and applying prevention principles-apply generally across

chronic diseases. As
well, there are disease specific aspects of self-management

which must be inte
grated with the general skills. These include particular

medications and specific
activities such as pursed-lip breathing, controlled

coughing and respiratory exer
cises for chronic lung disease, diets for heart disease and
special exercise for
arthritis. Such disease-specific activities are determined

together with the physi
cian or other health professionals. Once learned, they fit

well with general self
management skills. The Relevance of Perceived

Self-Efficacy to Self-Management and Health Outcomes in
Chronic Disease For patients to engage in effective
self-management, a number of pre
conditions are important. The first is an understanding of

the appropriateness
and value of the self-management activity. This

understanding flows from recog
nition of the nature of chronic illness and the

complementary roles of patients,
families and health professionals in management over time.

The second precon
dition is the development of skills and confidence on the
part of the patient
concerning the specific, useful self-management practices.

This is a matter of
learning, practicing and evaluating the personal benefits

derived from particular
self-management practices. The third precondition, highly

desirable but often
absent, is a health care service which encourages and

coordinates with self
management, preferably together with a social network which

supports and
facilitates self-management behaviors. The desired

preconditions generally fall outside of the classical
biomedical
model of health care in which the relatively passive

patient is attended by the all
knowing physician. Instead, the pre-conditions fall far

better under the biopsy
chosocial model of health and health care in which biology,

knowledge,
treatment, beliefs, emotions and socioeconomic

circumstances interact to
determine health outcomes (Engel, 1977). For instance,

cognitive and
motivational factors have long been recognized as

influencing the level of
functioning of patients with chronic disability; witness,

for example, the severely
deformed person with rheumatoid arthritis who leads an

independent and
satisfying life compared to persons with mild forms of the

same disease who
remain despondent and functionally incapacitated. Of

special relevance to self-management is the impact on
health status of
people's beliefs in their efficacy to exercise some control

over conditions that
affect their lives (Bandura, 1986). A growing body of

research has identified
various processes---cognitive, motivational, affective and

physiological-through
which self-beliefs of efficacy exert their effects.

Perceived self-efficacy influenc
es what people chose to do, their motivation, their

perseverance in the face of dif
ficulty, the self-enhancing or self-hindering nature of

their thought patterns, and
their vulnerability to stress and depression. Indeed,

people's beliefs in their per
sonal efficacy has been found to influence outcomes in a

number of acute and
chronic illnesses, including the level of benefit they

receive from therapeutic
interventions. Thus perceived self-efficacy provides a

linking mechanism
between psychosocial factors and functional status.

Beliefs in personal efficacy can be strengthened in four
principal ways
(Bandura, 1986). The most powerful relies on guided mastery

experiences that
build coping capabilities. In essence, this involves

learning and practicing the
appropriate behaviors. This is best done by breaking the

desired behavior into
small, graded tasks which can be accomplished in a

relatively short time. Feed
back is important so that patients can see progress. Once a

component task is
accomplished, another is added until the whole behavior is

achieved. Thus, a
patient may start a walking program by walking one block

four times a week.
Gradually blocks are added until the ultimate goal of

walking a mile four times
weekly is achieved. The second approach to building

personal efficacy draws upon the power of
social modeling to convey skills and a coping orientation.

This is the experience
of observing others exercise the skills and gain the

benefits. When using models
in teaching it is important that the models be as much like

patients as possible. If
a 60-year-old patient with osteoarthritis is attempting to

learn exercise, the model
should also be an older person experiencing problems with

arthritis. Prominent
persons and young models leading exercise programs do not

serve as good exam
ples for persons unlike them. There are two general types
of models for persons
suffering a handicap. Supermodels are those who have

reached some great
achievement in the face of much adversity. An example is

the young cancer
patient who walks across the country with an artificial

limb. While such models
are inspirational, they do little for enhancing efficacy

because people don't
believe they can come close to a similar achievement.

Coping models are people
who have a problem but cope with it on a day-to-day basis.
They have good and
bad days, but on the whole, are able to lead full and
active lives. These models
enhance efficacy because patients are able to relate

realistically to them. Social persuasion provides a third
type of efficacy enhancing influence; suc
cessful efficacy builders do more than convey positive

approaches, they also
design and explain activities for others in ways that bring

success. Effective per
suasion usually involves urging learners to do just a

little more than they are pre
sently doing. For instance, instead of urging a patient to

lose 40 pounds, it is bet
ter to urge them to lose a pound this week. The goal is

realistic and can be seen
by the patient as reasonable. In building self-efficacy,

success leads to success. The final mode of influence is
aimed at reducing aversive physiological
reactions or interpreting them in less pathological ways.

An example of the
former for a person with chronic lung disease would be

learning how to accomp
lish a desired physical activity without undesirable

shortness of breath, perhaps
through use of assisting devices; an example of the latter

for a person with
arthritis or heart disease would be recognizing that the

appearance of pain does
not commonly mean worsening disease. Perceived

self-efficacy is specific to particular activities. That
is, one can
feel efficacious for walking on a level surface but not for

mounting many stairs.
Therefore, just as the appropriate self-management
activities result from the par
ticular consequences of the chronic illness for a given

individual, so the devel
opment of perceived self-efficacy must relate to those

particular self-management
activities. With this background, it is pertinent to

review some recent evidence which
supports a mediating role for perceived self-efficacy in

the response to, and treat
ment of, chronic diseases. Historically, one of the first

applications of self
efficacy concepts occurred in the treatment of phobias

(Bandura, 1986). An
example is snake phobias which prevent individuals from

walking in fields or
other places where they fear snakes might lurk. Rapid and
successful ways of
curing established phobias were devised using different

combinations of the four
principal ways mentioned above to modify the patient's

perceived self-efficacy.
Since then, measures to enhance perceived self-efficacy

have been applied with
very salutary effects in a number of disease settings,

some which are discussed
elsewhere in this book. These settings include recovery

from heart attacks,
improvement of lung function in chronic pulmonary disease,

pain reduction in a
variety of settings such as headaches and child birth,

changing eating and exer
Bandura, 1991). In all these settings, patients and their

families have learned new
practices and behaviors which minimize or eliminate

dysfunction and discomfort. We shall illustrate the
application of self-efficacy principles and practices to
chronic disease through a brief recounting of their use in

patients with chronic
arthritis. Chronic arthritis limits the mobility of joints

and ultimately the strength
of muscles and tendons which move those joints. One of the

principles of good
medical management is exercise to retain or improve

mobility and strength. Such
exercise is frequently painful and it is common for

patients to imagine that exer
cise will make the joint disease worse. Good management

therefore involves not
only illustrating the appropriate exercises but also aiding

the patient to under
stand that, despite the pain, exercise is beneficial and

that as strength improves
pain will decline. Some years ago, we decided that a health

education program
about exercise and other aspects of management for chronic

arthritis would be
beneficial to groups of patients. A program called the

Arthritis Self-Management
Program (ASMP) was designed to cover, in six two-hour

sessions, such subjects
as: current understanding of the nature of arthritis,

present therapies and how
they are used, interpretation of symptoms, communication

with physicians and
other health professionals, self-management skills such as

exercise and cognitive
pain control, and marshalling of community resources. The
program was tested
in randomized, prospective studies, the results of which

showed that, compared
to control persons, participants in the ASMP experienced

significant reductions
in pain, less depression, increased social and physical

activities and decreased use
of physician services (Lorig, Lubeck, Kraines, Seleznick &

Holman, 1985). The
pain reductions were 15 to 20% from baseline levels which

is similar to that
achieved by the milder medications used for arthritis. In

order to understand the true effects of the ASMP, it was
important to know
whether the people who improved were those who learned the
most and practiced
what was taught most frequently. Because of large numbers

of participants, it
was possible to correlate health outcomes with knowledge

gained and reported
use of the self-management practices which were taught.

Surprisingly, the results
showed very little correlation; the best was 0.14 between

exercise and pain reduc
tion (Lorig, Seleznick, Lubeck, Ung, Chastain, & Holman,

1989). In order to
understand this unexpected turn, structured and open-ended

interviews were con
ducted with ASMP participants. The results revealed a

fairly consistent pattern:
Persons who did well in ASMP generally did not believe that
their chronic ar
thritis had irretrievably damaged their lives and believed
that they could do
things to improve matters; persons who did not do well had

opposite views
(Lenker, Lorig, & Gallagher, 1984). These findings

suggested that patients' per
ceived self-efficacy to cope with the consequences of

arthritis was mediating the
outcomes of ASMP. An instrument (see Appendix) was

therefore designed to
measure such perceived self-efficacy. In four months

following start of the
ASMP, perceived self-efficacy grew 10% to 15%. Further,

whereas the highest
correlation between health outcomes and either knowledge

gain or practice of
taught behaviors had been 0.14, as seen in Table 2, the

correlations of perceived
self-efficacy with health outcome scores were much higher

(Lorig, Chastain,
Ung, Shoor & Holman, 1989). As a consequence, the content

of ASMP was
modified to emphasize efficacy enhancing activities.

Subsequent outcomes of the
new ASMP were improved over those of the initial version

(Table 3). Recently, it has been possible to study two
groups of persons four years after
taking ASMP who had not experienced any organized

educational intervention in
the interim. Again surprisingly, as seen in Table 4, the

beneficial effects had
either persisted or increased. Pain remained reduced by

approximately 20% and
visits to physicians were reduced by approximately 40%.

Both of the latter
results occurred despite a modest increase in physical

disability. During the four
years, perceived self-efficacy to cope with the

consequences of arthritis as
measured by the final version of the instrument had grown

by an average of 29%
(Holman, Mazonson, & Lorig, 1989).
Table 2
Associations Between Self-Efficacy and Health Status at

Four Months After Start of the
Arthritis Self-Management Program (ASMP) Pain Depression

Self-efficacy for pain -0.39 a -0.45 Self-efficacy for
other symptoms -0.47 -0.59
Note. apearson correlations for combined data from 144

persons (49 from control group and 95 from ASMP). All
correlations significantly different from zero (p < 0.01).
Because of the persistent decline in use of physician
services over the four
years, it was possible to estimate cost savings which

accrued from the lesser
dependence on medical care. Taking into account the costs

of providing the
course and discounting the dollar value at 6% annually, it

was calculated that the
net four-year savings per participant with osteoarthritis

were approximately
$190.00 and the net savings per patient with rheumatoid

arthritis were almost
$650.00. Multiplying these savings by 1 % of individuals

suffering those two
diseases identifies potential savings approximating

$32,000,000 in four years if
only that small portion of patients participated in the

ASMP (or similar program)
and achieved the same results as have been found in the

initial studies. The health benefits for patients with
chronic arthritis described thus far have
been in the subjective (pain, depression) and behavioral

(activity, visits to physi
cians) realms. In one controlled study of persons with

moderately severe inflam
matory rheumatoid arthritis participating in the ASMP, pain

declined, self
efficacy grew and joint inflammation diminished over seven

weeks (O'Leary,
Shoor, Lorig, & Holman 1988). The latter suggested a

biological effect of the
ASMP on inflammation. However, measured immunological

parameters did not
change and the benefits were not sustained at four months.

Therefore, at this
time, there is no substantial evidence of a biological

benefit of the ASMP. The
results of ASMP and its evaluations indicate that health

education for self-man
agement, operating significantly through the vehicle of

perceived self-efficacy to
cope with the consequences of arthritis, results in

reductions in pain, depression
and dependence on medical care while improving

participants' physical and
social activities. These personal, social and financial

benefits strongly imply that
enhanced perceived self-efficacy is a vehicle through which

major advantage for
patients and for the health care system can be achieved

simultaneously.
Table 3
Comparison of Changes Achieved by the Original and Revised

Arthritis Self
Management Programs (ASMP)
Pain
O-to ScaJe
Disability
0-3 Scale
Depression
0-60 ScaJe
Note. *p < .15; ** P < .05.
Table 4 OriginaJ ASMP (N = 500) Baseline Four Month

Mean Change 4.8 -12% .7 -1% 12.7 -8% Revised ASMP (N
= 97) Baseline Four Month Mean Change 5.7 -\8%* .91
-10%** 15.2 -16%*
Summary of Changes in Outcome Attributes in Two Groups of

Participants at Four
Months and Four Years After Start of the Arthritis

Self-Management Course Group I Group II (N = 224) (N =
177)
Outcome Attribute Four Months Four Years Four Months Four

Years
Pain a -12% -19% -17% -22%
Disabilityb -3% +9% -4% +8%
Depression c -II% -2% -18% -3%
Visits to physicians -21% -43% -24% -44%
OriginaJ self-efficacy
measure +7% +17%
FinaJ self-efficacy
measure for pain +12% +34%
FinaJ self-efficacy
measure for other
symptoms +10% +25%
Note. aVisual AnaJogue ScaJe (O-to); bStanford Health

Assessment Questionnaire (Scale 0-3) (Fries, Spitz,
Kraines, & Holman, 1980); cBeck Depression Inventory
(0-39) (Beck, Ward, Mendelson, Mock, & Erbaugh, 1961).
The precise mechanism whereby perceived self-efficacy
exerts its effects on
patients with arthritis or other chronic diseases is not

known. The involved
patients have been heterogenous with diverse disease and

illness manifestations.
Hence it is not possible to infer that a specific

consequence of enhanced self
efficacy, such as handling snakes in the case of phobia,

mediated the overall
benefit. More likely, given the diversity of patients and

manifestations, indivi
dual patients developed understandings and adopted

practices which were
pertinent to their particular situation. Such a general

conclusion is consistent
with what is known about self-efficacy from many different

studies and settings.
Perceived self-efficacy affects people's willingness to

initiate change, the
magnitude of the change achieved, and their persistence in

the change. Because
self-efficacy is also specific to beliefs and activities,

it is reasonable that patients
identify areas of self-management pertinent to them and

develop a repertoire of
relevant attitudes and skills. In this regard, it is very

interesting that, in the
ASMP studies, self-efficacy grew, pain remained improved

and use of physician
services remained reduced over four years while physical

disability actually in
creased. A reasonable interpretation would hold that the

successful use of self
management skills reinforced and enhanced perceived

self-efficacy and sustained
the benefits despite a worsening physical situation.

Summary Chronic disease has replaced acute disease as the
predominant form of
disease in this country, and is the primary cause of

disability. Because chronic
disease can rarely be cured, its consequences extend out

over time. Sound man
agement of chronic disease and its consequences requires

participation of patients
and their families at most levels of health care from

understanding the disease to
applying management practices. Such participation, called

self-management,
requires a functioning partnership between the patient and

the physician. To
engage in effective self-management requires that the

patient achieve new knowl
edge and master new skills. In tum, the new learning and
skill development
appears to be dependent upon a person's perceived

self-efficacy in those realms.
Perceived self-efficacy can be developed and enhanced

through a variety of
learning experiences. High and increasing levels of
cope with the consequences of chronic disease are

associated with improvement
in symptoms, physical and emotional well-being and social

activities. Thus per
ceived self-efficacy is an essential precondition for the

appropriate management
of chronic disease. The present health care system was

developed in an era when acute disease
predominated, with concepts and structures adhering to the

attributes of acute
disease (Table 1). It can readily be argued that

significant responsibility for the
ineffectiveness and inefficiency in present health services

arises from the dis
cordance between a health care system designed to treat

acute disease and the
dominating prevalence of chronic disease. Designing health

services which are
more appropriate for chronic disease would be an essential

element in resolving
the health care crisis. The scope of such a design is

beyond the limits of this
chapter but, we argue, fostering and supporting

self-management practices and
developing the skills for them, would be an essential

ingredient of any successful

action: A social cognitive

promoting behavior. In 1. Madden (Ed.), Neurobiology of
Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., &

Erbaugh, 1. (1961). An inven
tory for measuring depression. Archives of General

Psychiatry, 4, 561-57l.
Colvez, A., & Blanchet, M. (1981). Disability trends in

the United States population
1966-76: Analysis of reported causes. American Journal of

Public Health, 7/, 464
471.
Engel, G. L. (1977). The need for a new medical model: A

challenge for biomedicine.
Science, /96, 129-136.
Fries, J. F., Spitz, P. W., Kraines, R. G., & Holman, H.

R. (1980). Measurement of
patient outcome in arthritis. Arthritis and Rheumatism, 23,

137-145.
Holman, H. R., Mazonson, P., & Lorig, K. (1989). Health

education for self-manage
ment has significant early and sustained benefits in

chronic arthritis. Transactions
Association of American Physicians, /02,204-208.
Holroyd, K. A., & Creer, T. L. (Eds.). (1986).

Self-management of chronic disease:
Handbook of clinical interventions and research. Athens,

OH: Academic Press.
Lenker, S., Lorig, K., & Gallagher, D. (1984). Reasons for

the lack of association
between changes in health behavior and improved health

status: An exploratory
study. Patient Education and Counseling, 6, 69-72.
Lewis, F. M., & Daltroy, L. H. (1990). How causal

explanations influence health
behavior: Attribution theory. In K. Glanz, F. M. Lewis, &

B. K. Rimer (Eds.), Health
behavior and health education theory research and practice

(pp. 90-114). San
Francisco, CA: Jossey-Bass.
Lorig, K., Chastain, R., Ung, E., Shoor, S., & Holman, H.
R. (1989). Development and
evaluation of a scale to measure the perceived

self-efficacy of people with arthritis.
Arthritis and Rheumatism, 32, 37-44.
Lorig, K., Lubeck, D., Kraines, R. G., Seleznick, M., &

Holman, H. R. (1985). Out
comes of self-help education for patients with arthritis.

Arthritis and Rheumatism,
28, 680-685.
Lorig, K., Seleznick, M., Lubeck, D. H., Ung, E.,

Chastain, R. L., & Holman, H. R.
(1989). The beneficial outcomes of the Arthritis

Self-Management Course are inade
quately explained by behavior change. Arthritis and

Rheumatism, 32,91-95.
O'Leary, A., Shoor, S., Lorig, K., Holman, H. (1988). A

cognitive-behavioral treat
ment for rheumatoid arthritis. Health Psychology, 7(6),

527-544.
Rice, D. P., & Feldman, J. J. (1983). Living longer in the

United States: Demographic
changes and health needs of the elderly. Milbank

Quarterly, 6/,362-395.
Rice, D. P., & LaPlante, M. P. (1988). The economics and
ethics of long-term care and
disability. Washington, DC: American Enterprise Institute

for Public Policy
Research.
Starr, P. (1982). The social transformation of American

medicine. New York: Basic
Books.
Verbrugge, L. M. (1984). Longer life but worsening health?

Trends in health and
mortality of middle-aged and older persons. Milbank

Quarterly, 62, 475-519. APPENDIX ARTHRITIS SELF-EFFICACY
SCALE
Self-Efficacy Pain Subscale
In the following questions, we'd like to know how your

arthritis pain affects you. For
each of the following questions, please circle the number

which corresponds to your
certainty that you can now perform the following tasks.
1. How certain are you that you can decrease your pain
quite a bit?
2. How certain are you that you can continue most of your
daily activities?
3. How certain are you that you can keep arthritis pain
from interfering with your sleep?
4. How certain are you that you can make a

small-to-moderate reduction in your
arthritis pain by using methods other that taking extra

medication?
5. How certain are you that you can make a large reduction
in your arthritis pain by
using methods other than taking extra medication?

Self-Efficacy Function Subscale
We would like to know how confident you are in performing

certain daily activities.
For each of the following questions, please circle the

number which corresponds to your
certainty that you can perform the tasks as of now,

without assistive devices or help
form another person. Please consider what you routinely can

do, not what would
require a single extraordinary effort.
AS OF NOW, HOW CERTAIN ARE YOU THAT YOU CAN:
l. Walk 100 feet on flat ground in 20 seconds?
2. Walk 10 steps downstairs in 7 seconds?
3. Get out of an armless chair quickly, without using your

hands for support?
4. Button and unbutton 3 medium-size buttons in a row in

12 seconds?
5. Cut 2 bite-size pieces of meat with a knife and fork in

8 seconds?
6. Tum an outdoor faucet all the way on and all the way off?
7. Scratch your upper back with both your right and left
hands?
8. Get in and out of the passenger side of a car without

assistance from another person
and without physical aids?
9. Put on a long-sleeve front-opening shirt or blouse

(without buttoning) in 8 seconds?
Self-Efficacy Other Symptoms Subscale
In the following questions, we'd like to know how you feel

about your ability to control
your arthritis. For each of the following questions,
please circle the number which
corresponds to the certainty that you can now perform the

following activities or tasks.
1. How certain are you that you can control your fatigue?
2. How certain are you that you can regulate your activity
so as to be active without
aggravating your arthritis?
3. How certain are you that you can do something to help

yourself feel better if you are
feeling blue?
4. As compared with other people with arthritis like yours,

how certain are you that you
can manage arthritis pain during your daily activities?
5. How certain are you that you can manage your arthritis
symptoms so that you can do
the things you enjoy doing?
6. How certain are that you can deal with the frustration
of arthritis?
Each question is followed by the scale: 10 20 very

uncertain 30 40 50 60 moderately uncertain 70 80 90
100 very certain
Each subscale is scored separately, by taking the mean of

the subscale items. The pain
and other symptoms scales can be scored separately or

together. The function scale
should not be combined with anything else. If one-fourth

or less of the data are
missing, the score is a mean of the completed data. If

more than one-fourth of the data
are missing, no score is calculated. (The authors invite

others to use the scale and
would appreciate being informed of the study results.)
This page intentionally left blank SELF -EFFICACY
EXPECTANCIES IN CHRONIC OBSTRUCTIVE PULMONARY DISEASE
REHABILITATION Michelle T. Toshima, Robert M. Kaplan, and
Andrew L. Ries The role of self-efficacy in the
rehabilitation of adult patients with chronic obstructive
pulmonary disease (COPD) was examined. One hundred and
nineteen COPD patients were randomly assigned to either a
comprehensive rehabilitation program or to an education
control group. Each program lasted two months. Patients
were evaluated on pulmonary function, exercise, treadmill
endurance, and psychosocial measures. These tests were
administered prior to the intervention, immediately
following the intervention, and one year after the start
of the program. The treadmill endurance walk and the
psychosocial measures were also administered six months
after the start of the program. Self-efficacy was
measured using a questionnaire that evaluated expectancies
to engage in specific activities that reflect the
functional disabilities often associated with COPD.
Validity of the self-efficacy construct was demonstrated
through systematic correlations with both pUlmonary
function and exercise variables. Although rehabilitation
patients demonstrated significant improvements in
treadmill performance, a trend toward improved
self-efficacy for walking was non-significant. Further,
the modest improvement in self-efficacy for walking did
not generalize to similar behaviors. Patients with high
initial self-efficacy scores for walking demonstrated the
greatest endurance on the treadmill. However self-efficacy
expectancies did not predict other health status outcomes.
We conclude that physiological feedback is a strong
source of self-efficacy expectation. These expectancies
might be modified by performance accomplishment. However,
continuing physiologic feedback provides a significant
obstacle for modifying self-efficacy in chronically ill
patients.
Recent advances in health psychology research have

demonstrated the impor
tance of cognitive variables in explaining and predicting

health behaviors. Social
learning theory and value-expectancy theory form the basis

for several cognitive
constructs that have been useful in explaining diverse

forms of health behavior
such as smoking cessation, pain management, and exercise
(see review O'Leary,
1985). Bandura's (1977) self-efficacy theory, in

particular, has received substan
tial empirical support for its explanatory role in the

therapeutic change process
and maintenance of treatment gains in smoking cessation

(Baer, Holt, & Lichten
stein, 1986; Devins & Edwards, 1988) the behavioral

treatment of pain (Dolce,
1987; Reese, 1982); exercise following uncomplicated

myocardial infarction
(Ewart, Taylor, Reese, & DeBusk, 1983); and exercise

training for patients with
chronic lung diseases (Kaplan, Atkins, & Reinsch, 1984).

Self-appraisal of capabilities can be altered in a number
of ways, including
mastery experiences, observing models, accepting social

persuasion, and altera
tion of physiological state (Bandura, 1977). The most

effective way of influenc
ing self-efficacy expectancies is through mastery

experiences (Bandura, 1977).
However, self-efficacy expectancies may develop through the

other channels
without direct experiences of a particular behavior. [n

this chapter, we consider
the importance of physiological feedback on self-efficacy

expectancies and sub
sequent task performance in patients with chronic

obstructive pulmonary disease
(COPD). Before proceeding, a brief overview of COPD and

the role of rehabili
tation in the management of patients with this disease will

be presented. CHRONIC OBSTRUCTIVE PULMONARY DISEASE AND
PULMONARY REHABILITATION RESEARCH Chronic obstructive
pulmonary disease (COPD) is a disorder characterized
by persistent expiratory airflow obstruction (American

Thoracic Society, 1987).
The diseases most often categorized as COPD include

emphysema, chronic bron
chitis, and irreversible asthma. Many patients exhibit

features of more than one
specific disease process. Although the etiology and

severity of these diseases
vary, the common clinical problem is impaired airflow which

results in the
symptom of shortness of breath. COPO is a major health

problem in the United States today; current estimates
reveal that 13.5 million Americans have COPD (Higgins,

1989). This condition
is the fifth leading cause of death in the U.S. and

accounts for approximately
71,000 deaths per year (National Center for Health

Statistics, 1988). In addition,
the death rate is increasing rapidly at a rate of 1.4% per

year, second only to
AIDS as the most rapidly increasing common cause of death

in the United States
(Lenfant, 1988). Since respiratory diseases are generally

considered to be of
greater importance as causes of morbidity and disability

than mortality, the eco
nomic consequences of COPO are great. COPD has been linked

to an estimated
4.7 million hospital days per year and 32.7 million

physician office visits per
year, approximately 5% of total physician visits (Feinleib

et al., 1989). Direct
and indirect costs for COPD were estimated to be $4.5

billion in 1972, $19 bil
lion in 1979, and $27 billion in 1982 (Lenfant, 1982).

Recent reports suggest that
1.14 million years of potential life are lost to COPO each

year (Kaplan, Atkins,
& Ries, 1985). The total mortality rate and years of

potential life lost rate are
higher for men than for women and higher for whites than
for other races. These
findings most likely reflect previous differences in

smoking patterns among these
groups (Morbidity and Mortality Weekly Report, 1986). COPD

has a profound impact upon the daily lives of afflicted
patients.
Dyspnea, the clinical symptom of shortness of breath,

interferes with daily activ
ities, often restricting patients to their homes. One of

the largest and most detail
ed studies on the quality of life of COPD patients was

reported by McSweeney
and colleagues (McSweeney, Grant, Heaton, Adams, & Timms,

1982). They
studied 203 patients suffering from COPD, and concluded

that these people were
significantly more impaired than a matched control group in

ambulation, self
care, social interaction, and recreational activities.

Depression and dissatisfaction
with life were also significantly more common in the COPD

patient. The progressive course of COPD is often
expressed in terms of functional
loss, impairment of gas exchange, and structural changes

in the lungs. For exam
ple, deterioration in forced expiratory volume in one

second (FEY I) for healthy
persons is estimated to be 20 to 30 mL/year. Deterioration

in FEY 1 for COPD
patients approximates 40 to 80 mL/year (Morbidity and

Mortality Weekly
Report, 1986). Lung dysfunction often causes symptoms that

result in a seden
tary lifestyle which further erodes the patient's

functional capacity. Thus, the
pulmonary patient often falls victim to a cyclical

downward pattern of increasing
disability and dyspnea. The patient who experiences

exertional dyspnea may
stop or restrict activities that produce the uncomfortable

and frightening symp
tom of dyspnea. Prolonged restriction of activities can

lead to the deconditioning
of diaphragmatic muscles and inefficient oxygen utilization

and ventilation. In
order to interrupt this cyclical increase in dyspnea and

the resulting decrease in
functional ability, pulmonary rehabilitation has been

promoted as an intervention
strategy (Bass, Whitcomb, & Forman, 1970; Mertens,

Shephard, & Kavanagh,
1978). A major component of pulmonary rehabilitation is a

structured exercise
program. In these programs, patients exercise on a

treadmill under close medical
supervision. Maximum exercise levels for COPD patients can

be predicted from
just a few variables (Carlson, Ries, & Kaplan, t 991) and
some patients can exer
cise at levels close to their maximum (Punzal, Ries,

Kaplan, & Prewitt, t 991).
However, since exercise causes shortness of breath, many

patients avoid physical
activity. Rehabilitation programs encourage regular, safe

exercise programs that
are generalized to home settings. Rehabilitation programs

for the management of COPD have expanded sub
stantially in recent years and continue to gain momentum.

The primary goal of
most programs is to restore the patient to the highest

possible level of function
ing. Results from previous studies suggest that

participation in a pulmonary re
habilitation program offers numerous positive outcomes for

COPD patients
(Dudley, Glaser, Jorgenson, & Logan, 1980; Moser,

Bokinsky, Savage,
Archibald, & Hansen, 1980). These benefits include

improved work tolerance
and work efficiency, improvement in activities of daily

living, reduced ventila
tory demand, decreased by dyspnea, and a reduction in

hospital days without any
appreciable change in pulmonary function. These outcomes

have been
documented using various research designs and training

programs. SELF -EFFICACY IN COPD REHABILITATION
Rehabilitation programs for COPD patients emphasize a
systematic increase
in activity levels through a structured exercise program. A

major component to
many exercise programs is treadmill walking in combination
with free walking.
The treadmill exercise experiences in rehabilitation

programs may affect self
efficacy expectancies through multiple sources of

infonnation. Many COPD
patients may have low self-efficacy because of their

preconceptions of pul
monary deficiencies and limitations. Years of experience

with a chronic illness
can result in negative symptoms such as shortness of

breath, fatigue, and pain.
These symptoms become more apparent when the patient

attempts activities such
as walking. The more prolonged or strenuous the activity,

the greater the in
crease in distressing symptoms. A strong preconception of

limitations and
impainnent increases the focus on negative physiological

reactions to exertion.
Therefore, it is likely that patients who focus on their

physical capabilities as they
engage in physically demanding tasks will judge their

pulmonary functioning as
stronger than those patients who selectively focus on the

discomforting sympto
matology. Providing patients with ongoing feedback about

their performance as
they encounter more challenging physical demands, can shift

the focus away
from the negative physiological feedback to the more

positive aspects of their
capabilities. Therefore, judgments of efficacy will vary

depending on how
patients interpret their symptoms during a physically
demanding task. Many patients with COPD often experience
moderate to severe discomfort
ing symptoms with even minimal exertion. For these

patients, their physiological
state may have an even greater impact on self-efficacy

judgments, particularly for
behaviors taxing the respiratory system. Self-efficacy

theory assumes that people
rely on inferences from their physiological state in

judging their capabilities.
Thus, for activities involving strength and stamina,

patients with COPD may
judge or interpret their shortness of breath, fatigue, and

pain as signs of physical
inefficacy. This chapter considers the validity of the

self-efficacy construct as well as its
function in promoting and maintaining change in patients

with COPD. In addi
tion, the association between sources of self-efficacy

infonnation, in particular
physiological feedback and activity levels, will be

evaluated in a clinical trial of
rehabilitation for patients with COPD. A test of the

validity of the self-efficacy
construct is provided by anchoring self-efficacy

expectancies, the hypothesized
mediator, to independently measurable variables of

pulmonary function such as
forced expiratory volume in one second (FEV 1)' residual

volume/total lung
capacity (RV{fLC), single breath diffusing capacity (OLCO),

exercise tolerance
as measured by oxygen uptake at maximum exercise (V0 2
max), and maximum
workload which can be estimated as metabolic equivalents

(METS max). Po stu
lated mediators, in this case self-efficacy expectancies,

are not directly observ
able. Nevertheless, theoretically self-efficacy

expectancies should be associated
with several observable indicators, in this case

physiological parameters of lung
functioning. This chapter considers self-efficacy

expectancies at two levels. We suggest
that self-efficacy in untreated patients is largely

determined by physiological
feedback. Thus, without intervention, the primary source

of information will be
disease severity. However, we also expect performance

accomplishment to over
ride physiological feedback. Specifically, patients who

participate in a rehabili
tation program designed to enhance activity should improve

in efficacy
expectation despite their physiological status. CLINICAL

TRIAL OF PULMONARY REHABILITATION Method
Subjects Over a one-year period, 350 patients with COPD

were screened for study;
129 met entry criteria and were randomized into either a

comprehensive pul
monary rehabilitation program or an education control

group. Ten patients drop
ped out prior to treatment, leaving 119 patients who

received the intervention.
There were no differences between those patients who

dropped out prior to the
intervention and those who remained in the study. The
subjects were 32 female
and 87 male patients. This female/male ratio approximates

the distribution of
COPD in females and males in the general population. In

order to be included,
the patient had to meet the following criteria: 1.

Clinical diagnosis of COPD, mild to severe, confirmed by
history, physical
examination, spirometry, arterial blood gases, and chest

roentgenograms.
Patients with emphysema, chronic bronchitis, or asthmatic

bronchitis were ac
cepted. Patients with primarily acute, reversible airway

disease (asthma) without
chronic airflow obstruction were not accepted. 2. Patients

were required to be stable on an acceptable medical
regimen. If
the treatment was considered inappropriate or the patient

was unstable, the pri
mary physician was contacted and the treatment regimen

adjusted prior to inclu
sion in the study. 3. Patients were excluded if they had

other significant disabling lung disease,
serious heart problems, or other medical conditions that

would interfere with
their participation. Assessment. Each patient underwent

pulmonary function tests, exercise tests,
treadmill endurance walks, and psychosocial measures prior

to the intervention
(baseline), immediately following the intervention (two

months) and one year
from the start of the program (twelve months). A six-month

assessment was also
conducted but involved only the treadmill endurance walk
and the administration
of the psychosocial measures. Due to the physical demands

of the pulmonary
function, exercise, and endurance walk tests, patients were

scheduled for the tests
on several occasions over a five-day period. Typically,

patients were given the
pulmonary function tests in the hospital laboratory on the

first day, the exercise
tests in the hospital laboratory on the third day, and the

treadmill endurance walk
and psychosocial measures in the rehabilitation building

on the fifth day. Certified cardiopulmonary respiratory
technicians administered the pul
monary function tests and exercise tests in the laboratory.

During the exercise
test, a pulmonary physician was also present to ensure the

safety of the patient.
The endurance walk tests and the psychosocial measures were

administered by
trained psychology graduate students who had current

cardio-pulmonary resusci
tation certification. In some instances where the patient

required continuous
electrocardiogram monitoring during the endurance walk

test, a pulmonary phy
sician and registered nurse observed for arrythmias and/or

premature ventricular
contractions. The physicians, technicians, and graduate

students were blind to
the group assignment of the patients.
Physiological Measures Descriptions of selected

physiological and psychosocial variables examined
in this chapter are presented below. Pulmonary junction

tests. Pulmonary function tests included: 1. Spirometry to
determine the following parameters: (a) Vital Capacity
(VC}--the maximum volume of air that can be expelled from

fully inflated
lungs; (b) Forced Expiratory Volume in one second (FEV I);

and (c) FEV INC
ratio. 2. Plethysmographic measurements: (a) Functional

Residual Capacity (FRC);
(b) Airway Resistance (RAW); (c) Residual Volume (RV); (d)

Total Lung
Capacity (TLC); and (e) RV/TLC ratio. 3. Single-breath

diffusing capacity for carbon monoxide (DLCO). 4. Maximal
inspiratory and expiratory pressures to assess respiratory
muscle
strength. Although pulmonary rehabilitation does not

typically lead to changes in
standard measures of pulmonary function, it was important

to monitor these
parameters in order to follow the progression of the

patient's disease. Exercise tests. The laboratory exercise
tests included: (a) an incremental,
symptom-limited exercise test to the maximal tolerable

level on a treadmill, and
(b) a treadmill test to define the steady-state walking

level for subsequent training
sessions. In the incremental exercise test, the work load

was increased at one
minute intervals by 0.5 miles per hour up to 3.0 miles per

hour with further work
increments made by increasing elevation by 2% to a maximum,

symptom-limited
level. This multiple-stage test assessed the maximal

exercise tolerance. In the
steady-state exercise test, the work load was maintained at
a constant level for a
predetermined period of time to allow the subject to reach

steady-state for the
variables of interest. This test was used to make

measurements at defined levels
for subsequent exercise training. After the incremental

exercise test and an appro
priate rest, patients performed the steady-state treadmill

test at the highest pos
sible level to determine a level for subsequent endurance

walk testing and
exercise training. During the tests, patients breathed

through a low-resistance breathing valve;
expired gases were analyzed continuously for measurements

of oxygen uptake
(V0 2 ), carbon dioxide elimination (VC0 2 ), expired

minute ventilation (V e)' and
other related variables. Metabolic equivalent (METS) was

estimated at the maxi
mal treadmill speed and grade as a measure of exercise

workload. During all
exercise tests an electrocardiogram with a single (modified

V 5) lead was used to
measure heart rate and monitor for arrhythmias or ischemia.

Blood pressure was
measured at regular three-minute intervals. Arterial blood

was sampled from an
indwelling radial artery catheter for measurement of

arterial oxygen pressure
(P a 0 2), arterial carbon dioxide pressure (p aC02), pH,

and alveolar-arterial
oxygen gradient (P(A-a)02) An ear oximeter was used to

monitor continuous
arterial oxygen saturation (Sa02). Patients who

demonstrated severe resting (P a02 < 55mmHg) or exercise
hypoxemia (P a02 < 50mmHg or Sa02 < 85%) were given

supplemental oxygen
and repeated the treadmHl exercise test on oxygen to

define a safe level for sub
sequent exercise training. Ratings of perceived

breathlessness and perceived
exertion were assessed after each exercise test using

standard scales adapted from
Borg (1982). Patients were asked to rate their degree of

breathlessness and
fatigue on a scale from 0 to 10, with 0 representing

nothing at all and 10 repre
senting the maximum ever experienced. Treadmill Endurance

Walk Test. Based on the maximal, symptom-limited
graded exercise test, each patient was given an exercise

prescription which
approximates maximal sustained exercise tolerance levels.

The target rates for
the individualized exercise prescriptions ranged from a

treadmill speed of 0.6
mph at 0% grade to 3.0 mph at 16% grade. The endurance

walk test was
designed to assess the patient's endurance for walking,

the type of exercise used
in the rehabilitation program. Prior to endurance walk

testing, heart rate, respi
ratory rate, and blood pressure were recorded with the

patient seated. Those
patients requiring supplemental oxygen waited for ten

minutes with the oxygen
prior to being tested. All patients walked at 1.0 mph for
two minutes (0.6 mph
for patients with that target speed). For patients whose

prescribed target rate was
higher than 1.0 mph, after two minutes the examiner asked
the patient if they felt
they could walk faster. If the patient replied yes or

maybe, the treadmill speed
was increased to a level 0.5 mph (or 2 to 4% grade) less

than the target
speed/grade. If the patient did not feel he or she could

walk faster, or if the target
rate was less than or equal to 1.0 mph, the initial speed
was maintained. After
two more minutes, if the target rate had not yet been
reached, the patient was
once again asked if he or she felt they could walk faster.

If the response was
positive, the treadmill speed was increased to the target

rate. Once the
individualized target rate was achieved, all patients were

instructed to walk as
long as possible. If the patient walked for 20 minutes at

the target rate, the
treadmill speed was increased another 0.5 mph (or 2 to 4%

grade). A maximum
endurance walk protocol was achieved if the patient walked

20 minutes at the
target rate and to minutes at a higher rate (i.e., 30

minutes maximum test).
During the entire endurance walk, the patient's blood

pressure was monitored
every three minutes. The test was stopped when a patient

stated that he/she was
unable to walk any longer. In addition, the examiner
stopped the test for any of
the following conditions: (a) chest pain; (b) dizziness;

(c) excessive rise in blood
pressure (> 250 mmHg systolic or > 130 mmHg diastolic

pressure); or (d)
excessive fall in blood pressure « 20 mmHg in systolic

pressure). At the
completion of the test, sitting pulse rate, respiratory

rate, and blood pressure were
measured. Symptoms of perceived breathlessness and fatigue

were rated after
the first two minutes on the treadmill and at the end of

the test, using a to-point
scale (Borg, 1982). The examiner then recorded the reasons

for stopping the test
(e.g., dyspnea, chest pain, maximum protocol).
Psychosocial Measures All patients completed a battery of

psychosocial measures at baseline, two
months, six months, and twelve months. The battery included

the following
measures: Self-Efficacy Questionnaire. The self-efficacy

questionnaire used in this
study was constructed and used in a previous study by

Kaplan, Atkins, and
Reinsch (1984) to demonstrate that specific rather than

generalized expectancies
mediate behavior changes in patients with COPO. The

self-efficacy question
naire was adapted from self-efficacy scales used to measure

levels of capability
to engage in activities that imposed stress on the heart

for patients with uncom
plicated myocardial infarction (Ewart et al., 1983). The
self-efficacy question
naire used in this study was modified to more accurately

measure the functional
disabilities associated with chronic obstructive pulmonary

disease. The question
naire consists of a list of seven behaviors that require

physical and/or emotional
stamina. The seven scales represent activities

progressively more dissimilar to
the target behavior of walking. Within each of the seven

scales is a series of brief
statements describing progressively more difficult

performance requirements.
For example, the scale for walking includes the following

statements: walk 1
block (approximately 5 minutes), walk 2 blocks (10

minutes), walk 3 blocks (15
minutes) ... walk 3 miles (90 minutes). The scale for

walking has nine items
representing unequal intervals of increasing difficulty.

For each item, the patient
rated the degree of confidence or strength of their

expectation to perform that
activity on a lOO-point probability scale, ranging in

to-point intervals from 0
(complete uncertainty) to 100 (complete certainty). The

seven scales included in
the measurement of self-efficacy expectancies are listed in

the Appendix. The
scale scores on the self-efficacy questionnaire reflect

the highest level that the
patient expressed 100% confidence they could perform or

tolerate the behavior. Quality 0/ Well-Being Scale. The
Quality of Well-Being (QWB) scale is a
comprehensive measure of health-related quality of life

that includes several
components. First, it obtains observable levels of

functioning at a point of time.
The levels of functioning are obtained from three separate

scales: mobility, phy
sical activity, and social activity. Second, symptomatic

complaints and disturb
ances are noted. Each patient is classified according to

the symptom or problem
that he or she finds most undesirable. Then, the observed

level of function and
subjective symptomatic complaint are weighted by preference

or the desirability
of the state on a scale ranging from 0 (for dead) to 1.0

(for optimum function).
The weights are obtained from independent samples of judges

who rate the
desirability of the observable health status. This system

has been used exten
sively in a variety of medical and health services

research applications (Kaplan &
Anderson, 1988). In addition, specific validity and

reliability studies using this
measure for COPD patients have been published (Kaplan,

Atkins, & Timms,
1984). These studies demonstrate that the QWB scale is

sensitive to relatively
minor changes in health status and that it is correlated

with a variety of physical
and functional measures of health status. Centers for

Epidemiologic Studies Depression Scale (CES-D). Depression
was measured using the CES-D scale. The CES-D scale is a
general measure of
depressive symptoms that has been used extensively in

epidemiologic studies
(Weissman, Sholomskas, Pottenger, Prusoff, & Locke, 1977).

The scale includes
twenty items and taps dimensions of depressed mood,

feelings of guilt and
worthlessness, appetite loss, sleep disturbance, and energy

level. These items are
assumed to represent all of the major components of

depressive symptomatology.
Sixteen of the symptoms are worded negatively, while the

other four are worded
positively to avoid the possibility of patterned responses.

The patient is asked to
report how often they experienced a particular "symptom"

during the past week
on a four-point scale: 0 (rarely or none o/the time-less

than J day), I (some or
a little o/the time-J to 2 days), 2 (occasionally or a

moderate amount o/time
3 to 4 days), 3 (most or all 0/ the time-5 to 7 days). The

responses to the four
positive items are reverse scored and then the total sum of
the responses is deriv
ed. Scores on the CES-D scale can range from 0 to 60 with

scores greater than
18 suggestive of clinically significant levels of

depression. The CES-D scale has
been found to have high internal consistency and

test-retest reliability (Radloff,
1977). It has also been documented to be highly correlated

to other standardized
depression scales (Weissman et aI., 1977). Eaton and

Kessler (1981) have
presented evidence for the reliability and validity of this

measure.
Group Comparisons of Baseline Measures As described in the

previous section, the patients were randomly assigned to
either a rehabilitation or an education contro group.

Patients in the rehabilitation
group participated in an outpatient rehabilitation program

consisting of education
about their disease, physical and respiratory care

instruction, psychosocial sup
port, and supervised exercise training, while patients in

the education control
group received infonnation about their disease through a

series of videotape and
lecture presentations. There were no significant

differences between patients in the rehabilitation or
education control group on baseline measures of

self-efficacy ratings, treadmill
perfonnance, and other psychosocial measures including

Quality of Well-Being
score, number of illness symptoms endorsed, depression,

and subjective ratings
of fatigue and dyspnea during the treadmill endurance walk.

These group com
parisons are presented in Table I.
Table 1
Group Comparisons of Variables at Initial Evaluation

Rehabili tation Education
Variable Mean SD Mean SD F

FEV! 1.39 0.66 1.44 0.62 1.12
RVrrLC 60.01 10.80 61.22 10.44 1.07
OLCO 14.17 7045 14.01 6.83 1.19
V0 2 max 1.24 0.51 1.24 0.54 1.12
Treadmill
endurance 12.37 8.36 11.79 7.97 1.10
QWB 0.6656 0.0960 0.6523 0.067 2.08
CES-O 14.02 8.74 15.34 10.03 1.32
Perceived
dyspnea 0049 2.19 4.52 2.09 1.10
Perceived
fatigue 4.28 2.23 4.25 2.17 1.06
SE-Walk 3.70 3.22 4.11 3.32 1.06
SE-Climb 2.11 lAO 2.08 1.51 1.16
SE-Lift 4.07 2.93 3.85 2.96 1.02
SE-Exert 2.30 1.16 1.94 1.25 1.16
SE-Push 2.81 1.14 2.55 1.33 1.35
SE-Stress 2.52 1.80 2.35 1.84 1.04
SE-Anger7 2.51 1.88 2.08 1.80 1.10
Note. SE == self-efficacy. Results Part i: Efficacy and

Physiological Feedback in Untreated Patients In order to
evaluate the relationship of baseline self-efficacy
judgments with
baseline physiological and psychosocial parameters, the

self-efficacy scores of all
patients were partitioned into tertiles, reflecting those

patients with the lowest,
middle and highest scores. A series of analyses of

variance tests were then
conducted. In these analyses, the efficacy tertiles for

walking, climbing, lifting,
pushing, exertion, stress, and anger served as independent

variables. A selected
group of pulmonary function, exercise, and psychosocial

assessment variables
were the dependent variables. The pulmonary function test

variables (including
FEV 1, RV{fLC, and OLCO) and the exercise test variables

(including
METS max , V0 2 max, and treadmill endurance) were

selected as representative
measures of lung disease severity and exercise tolerance,

respectively. Quality of
Well-Being and depression scores were the psychosocial

variables selected. The analyses suggest that
self-efficacy expectancies are significantly influ
enced by physiological state, particularly if the patient

has had little or no
mastery experiences with the task. The univariate F values

for the associations
between the pulmonary function test, exercise test, and

psychosocial variables
and self-efficacy expectancies for each category are shown

in Table 2. Figure I
summarizes the relationship between pulmonary function test

variables, FEV 1,
RV(fLC, and OLCO, and self-efficacy expectancies for

patients by lowest,
middle, and highest tertiles for each efficacy category.

As predicted, the self
efficacy measures reflecting a physical demand or task

(e.g., walk, climb, lift,
push, exert) were significantly related to the pulmonary
function variables. The
only exception was the non-significant relationship between

RV(fLC and self
efficacy expectancies for the push category, though the

trend was in the expected
direction. Similarly, Figure 2 summarizes the relationship

between exercise test
variables, V0 2 max and treadmill endurance, and tertiles

of self-efficacy expect
ancies for each efficacy category. As demonstrated with the

pulmonary function
measures, the exercise variables were also significantly

related to the self
efficacy measures reflecting a physical demand or task.

Thus, there was a highly
significant linear relationship between the physical

self-efficacy measures (e.g.,
walk, climb, lift, exert, push) and the pulmonary function

and exercise test var
iables. The relationship between the pulmonary function and

exercise test
measures and the self-efficacy judgments for stress and

anger, on the other hand,
were not systematically related. In contrast to the

systematic relationships observed with the pulmonary func
tion and exercise test variables, the relationships between

the psychosocial vari
ables, QWB, CES-O, and Treadmill, and tertiles of

self-efficacy expectancies for
each efficacy category were not as systematic. Figure 3

summarizes the relation
ship between psychosocial variables and self-efficacy

tertiles for each efficacy
category. The Quality of Well-Being scale much like the

psychological and exer
cise test variables was significantly related to tertiles

of self-efficacy involving a
physical demand. The CES-O scale, a scale in which higher

scores indicate more
depression, had significant linear relationships with

self-efficacy categories
measuring both a physical and emotional demand. It was not

significantly related
to self-efficacy for lifting, pushing, or tolerating

stress. Interestingly, the self
efficacy measures reflecting an emotional demand, in this

case, stress and anger,
were not systematically related to any of the pulmonary

function or exercise test
variables, nor to most of the psychosocial variables. 2

.... ;.. ~ ~ eJ, :t S 8 -= FEV 1 lly Self Efficacy
Category and TerliJe 2.0 1.5 1.0 Walk Climb Lift Exert
Push Anger Stress RvrrLC lly Self Eflicacy Category and
Tertilc 65 60 55 50 Walk Climb Lift Exert Push
AngerStress DLCO by Self Efficacy Category and Tertile ]
15 ~ o U ...l Q 10 Walk Climb Lift Exert Push Anger
Stress M. T. Toshima et aI. Lowest Middle Highest
Lowest Middle Highest Lowest Middle Highest
Figure 1 Relationship between pulmonary function measures

(FEV l' RV{fLC, & OLCO) and self-efficacy. 65 15 R V /
T L C Treadmill by Self Efficacy Category and Tertile 15
Lowost 10 Middle Highest 5 o Walk Climb Lift Exert
Push Anger Stress Efficacy V02max by Self Efficacy
Category and Tertile 1.5 Lowest Middle Highost 1.0
0.5 Walk Climb Lift Exert Push Anger Stress
Figure 2 Relationship between exercise test variables

(treadmill endurance & V0 2 MAX) with self-efficacy
tertiles for each efficacy category. 15 V O 2 m a x ( L /
m i n ) T r e a d m i l l ( m i n ) 1.5
Table 2
Analysis of Variance Values for Pulmonary Function Test,
Exercise Test, and
Psychosocial Variables by Self-Efficacy Tertilesfor Each

Efficacy Category Pulmonary Function Test Variables
Efficacy RV{fLC FEV) OLCO F P F P F P Walk 9.96
0.01 8.13 om 10.80 0.01 Climb 10.72 0.01 9.19 0.01 13.20
0.01 Lift 3.02 0.05 4.90 0.01 10.73 0.01 Exert 6.55
0.01 3.38 0.04 2.94 0.06 Push 2.32 n.s. 1.54 n.s. 2.83
n.s. Anger 2.32 n.s. 1.73 n.s. 1.78 n.s. Stress 1.59
n.s. 1.63 n.s. .10 n.s. Exercise Test Variables
Efficacy V0 2 max Treadmill F p F P Walk 13.11 0.01
10.11 0.01 Climb 18.74 0.01 7.99 0.01 Lift 15.99 0.01
2.29 n.s. Exert 3.64 0.03 5.08 0.01 Push 5.25 0.01 4.47
0.01 Anger 1.52 n.s. 3.12 0.05 Stress 0.57 n.s. 1.13
n.s. Psychosocial Variables Efficacy QWB CES-D F P F
P Walk 4.59 0.01 2.67 0.07 Climb 7.61 0.01 3.36 0.04
Lift 3.14 0.05 1.94 n.s. Exert 0.29 0.01 4.97 0.01
Push 11.78 0.01 0.15 n.s. Anger 0.46 n.s. 7.40 0.01
Stress 1.59 n.s. 1.54 n.s. 339 Qwn by Selr Erfieaey
Category and Terlile 0.725 0.700 0.675 Lowest Middle
0.650 Highest 0.625 0.600 Walk Climb Lift Exert Push
Anger Stress Erfieaey CES-D by SelrEffieaey Category and
Tertile 18 16 Lowes! M,ddle 14 Highest 12 10 Walk
Climb Lift Exert Push Anger Stress Erfieaey
Figure 3 Relationship between psychosocial variables (QWB,

CES-D) and selfefficacy for each efficacy category. 18 C
E S D Q W B Although self-efficacy expectancies are not
directly observable, the results
suggest that the efficacy expectancies are anchored to

physiological variables
which, in tum, are associated with observable levels of

behavior. Thus, the self
efficacy expectancies have observable correlates other than

the behaviors they
presumably govern. These data provide evidence for the

validity of the self
efficacy construct. Validity inferences are supported by

the anchoring of self
efficacy expectancies to independently measurable

indicators, in this case, physi
ological and exercise parameters. The results confirm that
external factors are
indeed linked to efficacy expectancies which, in tum, may

be linked to
observable behaviors. The degree and nature of the

relationship between self-efficacy expectancies
and behavior can be further quantified. Correlations

between aggregate self
efficacy scores and pulmonary function test variables,

exercise test variables, and
psychosocial variables were computed. Results of the

correlational analyses
revealed that the pulmonary function test, and the exercise

test were significantly
correlated with most of the aggregate perceived

self-efficacy scores. Table 3 dis
plays the relationship between aggregate efficacy

expectation scores for each
category and pulmonary function test and exercise test,

and psychosocial vari
ables respectively. High scores on RV(fLC and CES-D

reflect diagnostically
poorer status. Thus, negative relationships of these

variables and efficacy
expectancies were expected.
Table 3
Correlation Among Mean Pulmonary Function Test, Exercise

Test, and Psychosocial
Variables and Aggregate Self-Efficacy Scores for Each

Efficacy Category Self-Efficacy Category Walk Climb Lift
Exert Push Stress Anger
FEY1 .42** .35** .17* .18* .11 .04 .04
RY/fLC -.44** -.39** -.28** -.17* -.15* -.08 -.08

DLCO .46** .37** .18** .34** .17* .05 .06
Y02max .50** .41 ** .15** .45** .28** .02 .03
Treadmill
Endurance .42** .35** .27** .19** .27** .17* .18*
QWB .31 ** .42** .31** .18* .42** .09 .09
CES-D -.24** -.19* -.27** -.07 -.10 -.20* -.24**
Note. *p < .05; **p < .01. Although the magnitude of many
of the associations are significant, of most
interest is the nature of the relationships between the

variables and the efficacy
categories. For all physiological variables, the efficacy

category for walking was
most highly correlated, closely followed by climbing

and/or lifting. Moreover,
stress and anger efficacy categories were consistently

least correlated with the
physiological measures. In contrast, the pattern of

associations between the psy
chosocial variables and efficacy expectancies for the

various efficacy categories
did not follow the pattern of association found between

physiological variables
and efficacy expectancies. Self-efficacy for walking,

climbing, and lifting were
not necessarily the most highly correlated with the

psychosocial measures. In
addition, self-efficacy ratings for stress and anger were

not consistently the least
correlated with the psychosocial measures. Part 2: The

Effects of Performance Accomplishment Upon Efficacy
Expectations. Numerous studies in the health behavior
change literature suggest that self
efficacy expectancies play a crucial role in the initiation
and maintenance of
behavior change. Uniformly across studies of change in

health-related behavior,
including smoking cessation, pain management, and exercise

compliance, post
treatment self-efficacy judgments have been found to be

associated with out
come. In general, enhanced self-efficacy perceptions have

been associated with
greater behavior change and maintenance of gains over time.

Despite the sub
stantial literature base supporting this general finding,

only a select few have
examined self-efficacy expectancies across groups receiving

different health
behavior change interventions (Blittner, Goldberg, &

Merbaum, 1978; Chambliss
& Murray, 1979; Kaplan, Atkins, & Reinsch, 1984; Nicki,

Remington, &
MacDonald, 1985; Reese, 1982). This investigation not only

addresses the role
of self-efficacy expectancies on outcome, but addresses

changes in self-efficacy
judgments as a result of participation in a rehabilitation

program. Efficacy expectancies are based on four major
sources of information: mas
tery experiences, observing models, accepting social

persuasion, and alteration of
physiological state (Bandura, 1977). Successful experiences

performing a behav
ior (performance accomplishment) are the most potent

influence on self-efficacy
expectancies. In this investigation, it was hypothesized
that the rehabilitation
patients would have successful experiences walking on the

treadmill, as well as,
free walking on their own; thus, raising their efficacy

expectancies for walking.
Once efficacy expectancies for walking are established, it

was predicted that
enhanced efficacy for walking would generalize to other

similar behaviors as
demonstrated in a study by Kaplan, Atkins, and Reinsch

(1984). Thus, it was
predicted that improvements in walking behavior, and

subsequent increases in
self-efficacy expectancies for walking, transfer to similar

behaviors, and, to a les
ser degree, to behaviors dissimilar to those on which the

intervention was based. The effect of the rehabilitation
program on efficacy expectancies was evalu
ated in several ways. First, it was predicted that for

patients in the rehabilitation
group, self-efficacy expectancies for walking would

significantly improve fol
lowing participation in the rehabilitation program, while

no increases were ex
pected for the education control group. Secondly, for

patients in the rehabilita
tion group, it was proposed that improvements in

self-efficacy expectancies for
walking would generalize to efficacy expectancies for other

behaviors, with tasks
most similar to walking expected to have greater increases

in self-efficacy ex
pectancies than tasks dissimilar to walking. This gradient

of generalizability was
not expected to emerge in the education control group.

Self-efficacy expect
ancies were also predicted to be associated with health

status outcome measures. For both groups, strict
attendance records were kept to ensure that all patients
completed their respective programs. All patients were

required to make up any
sessions they had missed. Table 4 shows the status of

patient participation at the
various assessment periods. Two months after enrollment,

104 patients had
complete follow-ups, 4 patients had partial follow-ups, and

11 patients were un
available for testing. There was a 91 % follow-up rate at

two months, 89% at six
months, and 79% at twelve months. Attrition rates for the

two groups were not
significantly different, but by the twelve-month

assessment, twice as many edu
cation control patients than rehabilitation patients were

unavailable for follow-up.
Table 4
Patient Follow-Up Status by Group at Baseline. Two-Month.

Six-Month. and Twelve
Month Evaluations Number of Patients Baseline 2 Months 6

Months 12 Months
Status R E R E R E R E
Complete 57 62 48 56 50 53 46 42
follow-up
Partial 0 0 3 2 3 3
follow-up
No follow-up 0 0 6 5 6 7 8 17
(totals)
Deaths 0 0 0 0 1 0 2 2
Drop 0 0 1 1 2 4 2 5
Other 0 0 5 4 3 3 4 10
% of patients
tested 100 91 89 79
Note. R = rehabilitation, E = education control. At both

two-month and six-month post-intervention assessments,
patients in
the rehabilitation group performed significantly better on

the treadmill endurance
walk than education control patients as reported by

Toshima, Kaplan, and Ries
(1990). Furthermore, analysis of twelve-month follow-up

data demonstrate
continued superior performance on the endurance walk by the

rehabilitation
patients, though the effect was not significant. These

effects are shown
graphically in Figure 4. Treadmill perrormance by group

and roll ow-up period 22 20 18 16 14 12 -1 0 1 2 3 4
5 6 7 8 9 10111213 Month or follow-up Rehab Education
Figure 4 Comparison of treadmill endurance for

rehabilitation and education patients over 12 months.
Despite significantly improved treadmill performance by the
rehabilitation
patients at two-, six-, and twelve-month follow-up visits,

self-efficacy rating
scores for walking were not statistically significant

between the patient groups at
any of the three post-intervention assessments. Figure 5

shows self-efficacy rat
ing scores for both groups from baseline through

twelve-month assessments.
Although the self-efficacy ratings between groups were not

statistically signifi
cant, the pattern of scores was in the predicted direction.

There were significant differences between groups for
subjective ratings of
fatigue and dyspnea that were measured during the treadmill

endurance walk.
Not only did the patients in the rehabilitation group walk

longer and report some
what increased self-efficacy in their ability to do so,

they also reported less
fatigue and dyspnea at the end of the endurance walk test.

Repeated measures
analysis of variance for ratings of fatigue revealed

significant main effects for
group [F(I,67) = 5.21, P < .05) and time [F(1,67) = 5.73, p

< .01) and a signifi
cant group by time interaction [F(3,201) = 3.15, p < .051.

For ratings of dyspnea,
repeated measures analysis of variance revealed a

significant main effect for time
[F(I,67) = 7.05, P < .01] and a significant group by time

interaction [F(3,20l) =
6.77, p < .01). These differences in subjective symptom

ratings were statistically
significant at two-month follow-up for dyspnea [P(l,I04) =

10.69, p < .01) and
for fatigue [F(1,I04) = 10.05, p < .01). At the six-month

follow-up, the differ
ences were still significant for dyspnea [F(1,103) = 7.34,

p < .01] and for fatigue 22 T r e a d m i l l ( m i n . )
12
F(1,103) = 9.61, p < .01 J. By the twelve-month assessment,

the differences had
diminished somewhat, though still present. These results

are shown in Figure 6. Self efficacy for walking by group
and follow·up period 5.0 4.5 4.0 Rehab Education 3.5
·1 0 2345678910111213 Month of Follow.up
Figure 5 Comparison of walking efficacy expectations for

rehabilitation and education patients over 12 months.
Post·treadmill dyspnea and fatigue ratings by group and
follow.up period 5.5 5.0 4.5 4.0 3.5 3.0 2.5 ·1 0
2 3 4 5 6 7 8 9 10111213 Month of Follow-up Dyspnea(R)
Dyspnea(E) Fatigue(R) Fatigue(E)
Figure 6 Comparison of rating of perceived fatigue and

dyspnea for rehabilitation and education patients over 12
months. 5.0 R a t i n g s ( m i n 0 , m a x 1 0 ) S e l f
e f f i c a c y s c o r e f o r w a l k i n g 1 1
Correlations between self-efficacy for walking and
treadmill performance for
both rehabilitation and education control patients were

also assessed. For the
rehabilitation patients, correlations between the two

variables at each assessment
period are as follows: baseline (r = .42, P < .01), two

month (r = .23, n.s.), six
month (r = .39, p < .01), twelve month (r = .40, p < .01).

For the education con
trol patients, correlations between the two variables are

as follows: baseline (r =
.43, P < .01), two month (r = .30, P < .05), six month (r =
.32, P < .05), twelve
month (r = .30, n.s.). These findings provide further

support that patients in both
groups were fairly accurate in judging efficacy

expectancies for walking in rela
tion to actual treadmill performance. It had been

predicted there would be a systematic gradient for changes
in
efficacy expectancies for the rehabilitation but not for

the control group. Specifi
cally, the rehabilitation patients were expected to show

the largest change in
expectancies to perform the target behavior, walking, with

changes in expectan
cies for other behaviors changing as a function of their

similarity to walking.
However, the proposed gradient of generalizability did not

emerge and no signif
icant differences between the rehabilitation and education

control group were
noted for self-efficacy ratings of climbing, lifting,

exertion, pushing, anger
tolerance, or stress tolerance at any follow-up period. To

determine whether health status outcomes as measured by
Quality of
Well-Being and CES-D could be predicted from self-efficacy

expectancies and
other variables, multiple regression analysis was employed.

Both baseline and
post-intervention self-efficacy expectancies did not

account for a significant
amount of the variance in predicting Quality of Well-Being

or depression scores
at any of the assessment periods. Part 3: Microanalyses

of Self-efficacy and Behavior Change In a series of
experiments, Bandura, Reese, and Adams (1982) sought to
clar
ify the causal link between self-efficacy and behavior

through a method they
called microanalysis. This method examines the relationship

between differential
levels of efficacy expectancies and behavioral change.
Their first series of analy
ses demonstrated that performance varies systematically as

a function of perceiv
ed self-efficacy. Increasing levels of perceived

self-efficacy both across experi
mental groups and within experimental subjects resulted in

progressively higher
performance accomplishments. Thus, groups whose efficacy

expectancies were
raised to either low, medium, or high levels had

correspondingly low, medium, or
high performance attainment. Although there appears to be

a correspondence between enhanced self
efficacy expectancies and subsequent performance, efficacy

expectancies may
exceed, match, or remain below performance attainments,

depending on how
they are cognitively appraised. The following example

illustrates that self
efficacy expectancies are not merely a reflection of past

performance. Data from
several case study experiments (Bandura et al., 1982) show

how similar mastery
experiences have variable effects on perceived

self-efficacy over the course of
treatment. In one comparison, two moderately phobic

patients had very similar
performance attainment curves. In one case, self-efficacy

increased substantially
during initial successes but rapidly leveled off, even

though progressively more
demanding tasks were mastered, while the other patient

continued to exhibit a
steady increase in self-efficacy with each successive
mastery experience.
Because individuals are influenced more by how they

interpret their performance
successes than by the successes per se, perceived

self-efficacy is often a better
predictor of subsequent behavior than is past performance

attainments. Several analyses were conducted to evaluate
the role of differential levels of
initial perceived self-efficacy ratings on future

performance attainment. To facil
itate the analyses, baseline self-efficacy rating scores

for walking were categor
ized into tertiles with rating scores through one

comprising the low efficacy
group (n = 41), scores two through five comprising the

medium efficacy group
(n = 37); scores greater than five comprised the high

efficacy group (n = 41). To assess the relationship
between differential levels of initial self-efficacy
for walking and subsequent treadmill performance, an

analysis of variance across
groups for tertiles of efficacy for walking was conducted.

Rehabilitation and
education control patients were grouped together since

there were no initial dif
ferences between the groups on treadmill performance or

self-efficacy for walk
ing scores prior to the intervention. The results, as

displayed in Figure 7,
revealed a strong linear relationship between initial

perceived self-efficacy
expectancies for walking and treadmill performance [F(l,

119) = 10.11, p < .0 l[. Initial treadmill performance by
initial self efficacy for walking 16 14 12 10 8 Low
Medium High Efficacy for walking (tertile)
Figure 7 Treadmill endurance performance by initial

efficacy expectation for both groups combined. 16 8 T r
e a d m i l l ( m i n . ) A 2 x 3 (Group x Efficacy
tertiles) analysis of variance was conducted to
examine the role of differential levels of initial

self-efficacy expectations on
improvements in treadmill performance following the

intervention for both the
rehabilitation and education control patients. The results,

shown in Figure 8,
demonstrates significant main effects for group [F( I, I

08) = 26.98, P < .00 1\ and
self-efficacy for walking [F(2,106) = 2.86, p < .05], but

no significant inter
actions. The figure shows that for the rehabilitation

patients, those initially high
in self-efficacy for walking had less improvement in

treadmill performance,
while those initially low in self-efficacy for walking

demonstrated the greatest
improvements. For the education control patients, there was

little change in
treadmill performance regardless of initial self-efficacy

expectancies for walking. 11 Treadmill performance by
baseline efficacy for walking 14 12 10 8 6 4 2 o
Education Rehab Low Medium High Efficacy for Walking
(tertilr)
Figure 8 Initial efficacy expectations and subsequent

improvements in treadmill performance for rehabilitation
and education patients. DISCUSSION Patients with COPD
experience significant limitation in daily activities.
However, evidence from the pulmonary rehabilitation

literature suggests that
most of these patients can safely increase their levels of

exercise and activity.
Considering sources of information about self-efficacy, we

observed that psy
chological indicators of disease severity were

significantly correlated with
efficacy expectations. In other words, disease severity

was significantly asso
ciated with the expectation that patients could perform

various activities. Results
from Part I of the study suggest that without

intervention, physiological state 14 2 T r e a m i l l (
m i n . )
appears to dominate self-efficacy beliefs. Thus

physiological feedback, in the
fonn of symptoms, helps fonn expectations about

perfonnance. These findings
help define the validity of the self-efficacy construct and

suggest that it indeed
has meaning in the context of understanding function in

COPD patients. One of the problems in this line of
research is that physiological feedback
may not always be accurate. Although patients with COPD may

experience dis
comforting symptoms when active, most evidence suggests

that they can perfonn
more exercise without endangering their health. In fact,

exercise is typically
advocated as part of treatment. The intervention in this

investigation demonstrated that, despite these phy
siological limitations, significant increases in exercise

perfonnance were attain
able. Self-efficacy theory contends that mastery or

perfonnance-based success
experiences are the most potent influence on self-efficacy

expectancies. Thus, it
was predicted that patients in the rehabilitation program

would show increased
self-efficacy expectancies for walking as a function of

demonstrated improve
ments in actual treadmill perfonnance. Results from Part 2

of the study showed
improvements in exercise perfonnance. There are several

possible explanations for these findings. One explanation
for the non-significant results is that there was

insufficient statistical power to
detect a meaningful difference. With low power, true

differences might not be
detected. This study, however, was designed to have a .80

probability to detect
differences of .7 Z units. This might be classified as a

moderate effect size.
Thus, although moderate effects should have been detected,

small effect sizes
could have been missed. Another possible explanation is

that perfonnance accomplishment in this
group of COPD patients may not be as strong a source of

self-efficacy expectan
cies as might be predicted from self-efficacy theory. It

may be important to dis
tinguish between infonnation available to the individual

from the environment
and the individual's interpretation of the infonnation. The

impact of infonnation
on efficacy expectancies is dependent on how it is

cognitively appraised. Thus,
even success experiences do not necessarily create strong

generalized expectan
cies of personal efficacy. Success experiences are more
likely to produce posi
tive changes in efficacy expectancies if perfonnances are

perceived as resulting
from within the individual rather than from some external

source. The attribution
of increased perfonnance capabilities to external factors

such as the efforts of the
staff, rather than to the patients' own capabilities may

account for these findings. Numerous investigations of
health behavior change, have demonstrated
increased self-efficacy ratings for individuals who

participated in a variety of
intervention programs including smoking reduction and

cessation, pain manage
ment, and exercise maintenance. A majority of the studies,

however, made no
comparisons to a control group. In this study, the

rehabilitation patients certainly
gained in their ability to walk, but did not show an

equally strong increase in self
efficacy expectancies for walking, while the education

control patients did poorly
on the treadmill exercise and showed little change in

efficacy expectancies. Previous investigations have shown
that self-efficacy expectancies derived
from successful performance with a particular task

generalized along dimensions
of task similarity to other behaviors (Kaplan, Atkins, &

Reinsch, 1984). In this
investigation, the hypothesized gradient of

generalizability of self-efficacy
expectancies did not emerge. This finding was unexpected

because the study
used the same measure and a similar patient group. One
possible explanation for
the divergent finding is that the previous study used a

fairly specific, cognitive
behavioral intervention rather than a medical model

rehabilitation intervention as
employed in this investigation. The critical difference

being the cogniti ve behav
ioral intervention stressed internal attributions for

behavior change. Various studies have also shown
self-efficacy ratings to be predictive of suc
cessful performance with many responses including phobias

(Bandura, Adams,
Hardy, & Howells, 1980), physical stamina (Ewart et al.,

1983; Kaplan, Atkins,
& Reinsch, 1984; Weinberg, Gould, & Jackson, 1979) and

self-regulation of
addictive behaviors (Coelho, 1984; Condiotte &

Lichtenstein, 1981; Godding &
Glasgow, ] 985). The findings from this investigation

suggest that self-efficacy
expectancies for walking predict actual treadmill

perfonnance. However, self
efficacy judgments did not predict other health status

outcomes. This supports
self-efficacy theory and suggests that self-efficacy

expectancies are task specific
and do not necessarily reflect more general outcomes. The

finding that the rehabilitation patients reported
significantly decreased
dyspnea and fatigue following the treadmill endurance walk

is important.
Dyspnea is recognized in the literature as one of the most

disabling symptoms of
COPD. The literature addresses at length the fear/dyspnea
cycle. Basically, the
fear/dyspnea cycle refers to the downward cyclical process

whereby patients are
limited in their functional capabilities because of their

fear of dyspnea. This fear
of dyspnea often produces increased anxiety and leads to

greater feelings of
shortness of breath, further limiting the patients ability

to engage in activities. At
the present time, the interaction of environmental stimuli,

psychological var
iables, physical capacity, and physiological mechanisms on

the perception of
dyspnea is not well understood. Attempts to reduce the

symptom of dyspnea
through pharmacological and medical interventions have

produced mixed and
often conflicting results. The finding from this study

suggests that rehabilitation
may be an effective method to break the cycle by

decreasing the fear associated
with dyspnea, ultimately leading to greater functioning.

The findings from Part 3 of the study suggest that initial
beliefs about abil
ities to perform a certain behavior are important

predictors of subsequent
performance capabilities. Prior to the intervention, there

was a clear and direct
relationship between self-efficacy ratings for walking and

actual treadmill per
formance, with patients high in self-efficacy for walking

demonstrating the long
est duration on the treadmill. Conversely, those patients

low in self-efficacy had
the poorest performance on the treadmill. These results

are consistent with
previous investigations of health behavior change, and

indicate that self-efficacy
is a good predictor of perfonnance. Although initial

self-efficacy expectancies for walking predicted treadmill
perfonnance before the intervention, following the

intervention, initial self
efficacy expectancies had differential effects on

improvements in treadmill per
fonnance. It was predicted that higher efficacy

expectancies would predict
higher perfonnance capabilities. An unexpected finding was

that patients in the
rehabilitation group with initially low self-efficacy

expectancies showed the
greatest gains in treadmill perfonnance, while those

patients with initially high
self-efficacy expectancies had the smallest perfonnance

gains. One possible
explanation is that patients with initially low

self-efficacy had the greatest
improvement, since a perfonnance ceiling may have prevented

those highest in
efficacy at baseline from achieving even higher perfonnance

accomplishments.
In other words, the patients who started out at initially

low levels of efficacy for
walking had the greatest opportunity for making

improvements because they
started out at lower levels of performance. Another

possible explanation is that
the rehabilitation program may have produced some increases
in efficacy expect
ancies, and again those with initially low levels had the
most to gain. It it also
possible that self-efficacy expectancies have less value

for those who are already
perfonning at a high level. CONCLUSIONS Contributions to

Self-Efficacy Theory Measurement of self-efficacy.
According to self-efficacy theory, self-efficacy
expectancies are established through both physiological and

psychological feed
back. The theory suggests that feedback from physiological

indicators has a
more central role in establishing efficacy expectancies

when there has been little
or no exposure to the task or behavior. The results from

this investigation lend
support to this assertion and suggest that, indeed,

physiological variables, as
measured by pulmonary function and exercise tests, are

associated more strongly
with self-efficacy judgments than psychosocial variables.

These results demon
strate that self-efficacy expectancies are associated with

observable correlates
other than the behaviors they presumably govern; thus,

providing evidence for
the validity of the self-efficacy construct. This is one

of a handful of experiment
al investigations that has systematically evaluated the

validity of the construct. Self-efficacy enhancement
through success experiences. Self-efficacy theory
maintains that efficacy expectancies can be influenced by

various sources of
incoming infonnation, the most influential being
performance attainment. In this
investigation, rehabilitation resulted in significant

increases in perfonnance
attainment without similar significant increases in

self-efficacy. In comparison,
the control group experienced no improvements in treadmill

perfonnance
accomplishment or self-efficacy expectancies. Although

self-efficacy theory
maintains that perfonnance accomplishment is the most

influential factor in
enhancing self-efficacy expectancies, the data from this

study suggest that self
appraisals of perfonnance accomplishment may be attenuated

by a number of
contextual factors. For example, in this patient

population, efficacy changes
derived from perfonnance accomplishment may be attenuated

by strong physio
logical feedback. For this reason, even success experiences

do not necessarily
create strong expectancies of self-efficacy. Therefore,

although perfonnance
accomplishment is an important factor in building

efficacy, the infonnation from
successful perfonnance can be attenuated, leading to

smaller changes in self
efficacy expectancies than might be expected. The present

research findings do not provide unanimous support for
princi
ples of self-efficacy theory (Bandura, 1977, 1982, 1986).

Self-efficacy theory
suggests that while specific procedures for achieving

change may differ for
different clinic populations, the general strategy of

assessing and enhancing self
efficacy expectancies by providing perfonnance mastery

experiences has sub
stantial utility. The results from this investigation raise

some questions about the
functional relationship between successful perfonnance

experiences and
enhancement of efficacy expectancies. Although the value of

self-efficacy
expectancies in health promotion and maintenance has been

demonstrated in
numerous studies, the present results provide mixed support

for the usefulness of
the self-efficacy construct in understanding health

behavior change in COPD
patients. Further investigations into self-efficacy

expectancies in more seriously
ill patients is needed in elucidating the role

self-efficacy expectancies play in
American Thoracic Society. (1987). Standards for the

diagnosis and care of patients
with Chronic Obstructive Pulmonary Disease (COPD) and

asthma. American Review
of Respiratory Disease, J 36( I ), 225-244.
Baer, 1. S., Holt, C. S., & Lichtenstein, E. (1986).

Self-efficacy and smoking reexam
ined: Construct validity and clinical utility. Journal of

ConSUlting and Clinical

Bandura, A (1982). Self-efficacy mechanism in human

agency. American Psycholo
gist, 37, 122-147.
Bandura, A (1986). Scope of self-efficacy theory. Journal

of Social and Clinical
Bandura, A, Adams, N. E., Hardy, A 8., & Howells, O. N.

(1980). Tests of the gener
ality of self-efficacy theory. Cognitive Therapy and

Research, 4, 39-66.

Microanalysis of action and fear

Bass, H., Whitcomb, 1. F., & Forman, R. (1970). Exercise

training: Therapy for
patients with chronic obstructive pulmonary disease.

Chest, 57, 116-121.
Blittner, M., Goldberg, 1., & Merbaum, M. (1978).

Cognitive self-control factors in the
reduction of smoking behavior. Behavior Therapy, 9,

553-561.
Borg, G. A. V. (1982). Psychophysical bases of perceived

exertion. Medical Science
Sports Exercise, 14, 377-381.
Carlson, D.J., Ries, A.L., & Kaplan, R M., (in press).

Prediction of maximum excersise
tolerance in patients with chronic obstructive pulmonary

disease. Chest.
Chambliss, C. A., & Murray, E. 1. (1979). Cognitive
procedures for smoking reduction:
Symptom attributions versus efficacy attribution.

Cognitive Therapy and Research,
3,91-95.
Coelho, J. E. (1984). Self-efficacy and cessation of

smoking. Psychological Reports,
54, 309-310.
Condiotte, M. M., & Lichtenstein, E. (1981). Self-efficacy

and relapse in smoking ces
sation programs. Journal of Consulting and Clinical

Devins, G. M., & Edwards, P. J. (1988). Self-efficacy and

smoking reduction in chron
ic obstructive pulmonary disease. Behavior Research and

Therapy, 26, 127-135.
Dolce, 1. 1. (1987). Self-efficacy and disability beliefs

in behavioral treatment of pain.
Behavior Research and Therapy, 25, 289-299.
Dudley, D. L., Glaser, E. M., Jorgenson, B. N., & Logan,

D. L. (1980). Psychosocial
concomitants to rehabilitation in chronic obstructive

pulmonary disease: Part I. Psy
chosocial and psychological considerations; Part 2.

Psychosocial treatment; Part 3.
Dealing with psychiatric disease. Chest, 77,413-420;

544-551; 677-684.
Eaton, W. W., & Kessler, L. G. (1981). Rates of symptoms

of depression in a national
sample. American Journal of Epidemiology, 114,528-538.
Ewart, C. K., Taylor, C. B., Reese, L. B., & Debusk, R. F.

(1983). Effects of early post
myocardial infarction exercise testing on self-perception
and subsequent physical
activity. American Journal of Cardiology, 51, 1076-1080.
Feinleib, M., Rosenberg, H. M., Collins, 1. G., Delozier,

J. E., Pokras, R., & Chevarley,
F. M. (1989). Trends in COPD morbidity and mortality in the

United States. In:
The rise in chronic obstructive pulmonary disease

mortality. American Review of
Respiratory Diseases, 140 (suppl3 pt 2), S9-S 18.
Godding, P. R, & Glasgow, R. E. (1985). Self-efficacy and

outcome expectancy as

Higgins, M. W. (1989) Chronic airways disease in the

United States: Trends and
determinants. Chest, 96 (suppl), 328s-334s.
Kaplan, R. M., & Anderson, 1. P. (1988). A general health

policy model: Update and
applications. Health Services Research, 23,203-235.
Kaplan, R. M., Atkins, C. 1., & Reinsch, S. (1984).

Specific efficacy expectancies

Kaplan, R. M., Atkins, C. 1., & Ries, A. L. (1985).

Behavioral issues in the manage
ment of chronic obstructive pulmonary disease. Annals of

Behavioral Medicine, 7, 5
10.
Kaplan, R M., Atkins, C. J., & Timms, R M. (1984).

Validity of a quality of well
being scale as an outcome measure in chronic obstructive

pulmonary disease.
Journal of Chronic Diseases, 37,85-95.
Lenfant, C. (1982). Government and Community. American

Review of Respiratory
Disease, 126(15),753-757.
Lenfant, C. (1988). Introduction. In A. J. McSweeney & I.

Grant (Eds.), Chronic
Obstructive Pulmonary Disease: A behavioral perspective

(pp. iii-iv). New York:
Marcel Dekker.
McSweeney, A. 1., Grant, I., Heaton, R. K., Adams, K. M., &

Timms, R. M. (1982).
Life quality of patients with chronic obstructive

pulmonary disease. Archives of
Internal Medicine, 142,473-478.
Mertens, D. 1., Shephard, R. J., & Kavanagh, T. (1978).

Long-term exercise therapy for
chronic obstructive pulmonary disease. Respiration,

35,96-107.
Morbidity and Mortality Weekly Report (1986). Deaths due to

Chronic Obstructive
Pulmonary Disease and Allied Conditions. MMWR, 35(32),

507-5LO.
Moser, K. M., Bokinsky, G. E., Savage, R. T., Archibald, C.

1., & Hansen, P. R. (1980).
Results of a comprehensive rehabilitation program:

Physiologic and functional
effects on patients with Chronic Obstructive Pulmonary

Disease. Archives of Internal
Medicine, 140, 1596-1601.
National Center for Health Statistics. (1988).Current

estimates from the National Health
Interview Survey, United States. Vital and Health
Statistics. Series 10, No. 173,
DHHS Pub. No. (PHS)90-1232. Washington, DC: US Department

of Health and
Human Services, 1990.
Nicki, R. M., Remington, R. E., & MacDonald, G. A. (1985).

Self-efficacy, nicotine
fading/self-efficacy monitoring and cigarette smoking

behavior. Behavior Research
and Therapy, 22, 477-485.
O'Leary, A. (1985). Self-efficacy and health. Behavior

437-451.
Punzal, P. A., Ries, A. L., Kaplan, R. M., & Prewitt, L.

M. (in press). Maximum inten
sity exercise training in patients with chronic

obstructive pulmonary disease. Chest.
Radloff, L. S. (1977). The CES-D Scale: A self-report

depression scale for research in
the general population. Applied Psychological Measurement,

I, 385.
Reese, L. B. (1982). Pain reduction through cognitive,

self-relaxative and placebo
means: A self-efficacy analysis. Unpublished doctoral

dissertation, Stanford
University, Stanford, California.
Toshima, M. T., Kaplan, R. M., & Ries, A. L. (1990).

Experimental evaluation of reha
bilitation in chronic obstructive pulmonary disease:

Short-term effects on exercise
endurance and health status. Health Psychology, 9, 237-252.

Expectancies and performance-An

of Sports Psychology, I,
320-331.
Weissman, M. M., Sholomskas, D., Pottenger, M., Prusoff,

B. A., & Locke, B. Z.
(1977). Assessing depressive symptoms in five psychiatric

populations: A validation
study. American Journal of Epidemiology, 106,203-214.

Author Notes Supported by Grant ROl HL 34732 from the
National Heart, Lung, and
Blood Institute. Address requests for reprints to the

second author, Robert M.
Kaplan. APPENDIX SELF-EFFICACY EXPECTANCIES MEASURE
Instructions: The following measure describes various tasks

and activities. Under the
column marked Can Do, put a check mark next to the tasks or
activities you expect you
could do now. For each of the tasks you checked under Can
Do, indicate in the column
marked Conf., how confident you are that you could do that
task now. Rate the degree
of your confidence using a number from 0 to 100 on the

scale below: o 10 20 30 40 50 60 70 80 90 100 Uncertain
LIFrING OBJECfS
Lift a 10 lb. object

CLIMBING (without rest)
Walk up several stairs
Walk up I flight of stairs
Walk up 2 fiight of stairs
Walk up 3 flight of stairs
Walk up 4 flight of stairs
PUSHINGIMOVING OBJECfS
Move a light weight object
(Kitchen chair)
Move a medium weight
object (Coffee table)
Move a fairly heavy object
(Armchair)
Move a heavy object
(Sofa or Bed) Moderately Certain Certain WALKING

.ClIn...QQ Qmf. Walk I block (5 min.) Walk 2 blocks (10
min.) Walk 3 blocks (15 min.) Walk 4 blocks (20 min.)
Walk 5 blocks (25 min.) Walk 1 mile (30 min.) Walk 1.5
miles (45 min.) Walk 2 miles (60 min.) Walk 3 miles (90
min.) GENERAL EXERTION Capable of very light exertion
Capable of light exertion Capable of moderate exertion
Capable of hard exertion Capable of very hard exertion
Capable of extremely hard exertion TOLERANCE OF
EMOTIONAL TENSIONISTRESS Tolerate mild tension/stress
Tolerate some tension/stress Tolerate moderate
tension/stress Tolerate substantial tension/stress
Tolerate a great deal of tension/ stress TOLERANCE OF
ANGER AROUSAL Tolerate mild anger arousal Tolerate some
anger arousal Tolerate moderate anger arousal Tolerate
substantial anger arousal Tolerate a great deal of anger
arousal _ Can Do Conf. SELF·EFFICACY MECHANISM IN
PSYCHOBIOLOGIC FUNCTIONING Albert Bandura Perceived
self-efficacy operates as an important psychological
mechanism linking psychosocial influences to health
functioning. Perceived self-efficacy affects a wide range
of biological processes that mediate human health and
disease. Many of these biological effects arise in the
context of coping with acute and chronic stressors.
Exposure to stressors with a sense of efficacy to control
them has no adverse effects. But exposure to the same
stressors with perceived inefficacy to control them
activates autonomic, catecholamine, and opioid systems and
impairs the functioning of the immune system. Depending
on their nature, lifestyle habits enhance or impair health
status. This enables people to exercise some control over
their vitality, quality of health, and rate of aging.
Self-efficacy beliefs affect every phase of personal
change-whether people even consider changing their health
habits; whether they can enlist the motivation and
perserverance needed to succeed should they choose to do
so; and how well they maintain the changes they have
achieved. Health outcomes are related to predictive
factors in complex, multidetermined and probabilistic
ways. Prognostic judgments, therefore, involve some degree
of uncertainty. Because prognostications can alter
selfefficacy beliefs, such judgments have a self-validating
potential by influencing the course of health outcomes.
The recent years have witnessed a major change in the

conception of human
health and illness. The traditional approaches relied on a

biomedical model
which places heavy emphasis on infectious agents,

ameliorative medications, and
repair of physical impairments. The newer conceptions

adopt a broader bio
psychosocial model (Engel, 1977). Viewed from this

perspective, health and di
sease are products of interactions among psychosocial and

biological factors.
Health is not merely the absence of physical impairment and

disease. The bio
psychosocial perspective emphasizes health enhancement as

well as disease pre
vention. It is just as meaningful to speak of degrees of

wellness as of degrees of
impairment. Thus, for example, there are degrees of

immunocompetence, cardio
vascular robustness, physical strength and stamina,

movement flexibility and
cognitive functioning. Health enhancement seeks to raise

the level of
psychobiological competencies. It is now widely

acknowledged that people's health rests partly in their own
hands. In analyzing mortality rates within and between

countries, Fuchs (1974)
has shown that expenditures for medical care have only a

small impact on life
expectancy. The quality of health of a nation is largely

determined by lifestyle
habits, and environmental conditions. People often suffer

physical impairments,
and die prematurely of preventable health-impairing

habits. Industrial and agri
cultural practices are injecting carcinogens and harmful

pollutants into the air we
breathe, the food we eat, and the water we drink, all of

which take a heavy toll on
the body. Changing health habits, and environmental

practices yields the large
health benefits. Psychosocial determinants of health

status operate largely through the exer
cise of personal agency. Among the mechanisms of personal
agency, none is
more central or pervasive than people's beliefs in their

capability to exercise con
trol over their own motivation and behavior and over

environmental demands.
Evidence from diverse lines of research shows that

perceived efficacy operates as
an important psychological mechanism linking psychosocial

influences to health
functioning. One can distinguish two levels of research on

the psychosocial de
terminants of health functioning in which perceived

self-efficacy plays an influ
ential role. The more basic level examines how perceived

coping self-efficacy
affects biological systems that mediate health and disease.

The second level is
concerned with the exercise of direct control over the

modifiable aspects of
health and the rate of aging.
BIOLOGICAL EFFECTS OF PERCEIVED SELF -EFFICACY Perceived

self-efficacy can activate a wide range of biological
processes that
mediate human health and disease. Many of these biological

effects arise in the
context of coping with acute or chronic stressors in the

many transactions of
everyday life. Stress has been implicated as an important

contributing factor to
many physical dysfunctions (Goldberger & Breznitz, 1982;

Krantz, Grunberg, &
Baum, 1985). Recent investigations with animals have

identified controllability
as a key organizing principle regarding the nature of
stress effects. Exposure to
stressors with a concomitant ability to control them has no

adverse effects. How
ever, exposure to the same stressors without the ability to

control them activates
neuroendocrine, catecholamine, and opioid systems and

impairs the functioning
of the immune system (Bandura, 1991; Coe & Levine, 1991;

Maier,
Laudenslager, & Ryan, 1985; Shavit & Martin, 1987).

Biochemical Effects of Self-Efficacy in Coping With
Stressors Social cognitive theory views stress reactions
in terms of perceived self-inef
ficacy to exercise control over aversive threats and taxing

environmental de
mands. If people believe they can deal effectively with

potential environmental
stressors they are not perturbed by them. But if they

believe they cannot control
aversive events they distress themselves and impair their

level of functioning.
Our understanding of the biological effects of

uncontrollable stressors is based
mainly on experimentation with animals involving

uncontrollable physical stres
sors. Stressors take diverse forms and can produce

different patterns of physio
logical activation. This places certain limitations on

extrapolation of conclusions
across different species, stressors, and patterns of

controllability. Uncontrollable
physical stressors are not only stressful but also inflict

some physical trauma that
can activate a variety of complicating physiological
processes. Most of the im
portant stressors with which humans have to cope involve

psychological threats
(Lazarus & Folkman, 1984). Moreover, stress reactions are

governed largely by
perception of coping self-efficacy rather than being

triggered directly by the ob
jective properties of threats and environmental demands

(Bandura, 1988a). It is
the perception of environmental threats as exceeding one's

coping capabilities
that becomes the stressful reality. Efforts to verify the

effects of controlling efficacy on biological stress
reactions in humans have relied extensively on

correlational or quasi-experi
mental studies in which occurrences of life stressors are

related to indices of
biological functioning or infectious illnesses. Such

studies leave some ambiguity
about the direction of causality and even whether the

biological effects are due to
the stressor or other unsuspected factors operating at the

time. To overcome
these problems, we devised a research paradigm combining

strong phobic stres
sors with mastery efficacy induction procedures that

enables us to examine causal
relationships under laboratory conditions with a high

degree of experimental
control over confounding sources of influence.

Participants cope with a uniform
stressor that can be varied in intensity. Because a high

sense of controlling effi
cacy can be quickly instilled through guided mastery

experiences, we can create
conditions combining exposure to chronic stressors with,

and without, perceived
controlling self-efficacy. By the end of each study, the

phobia is eradicated in all
participants so they all gain lasting relief from chronic

phobic stressors while
contributing to knowledge. Autonomic Activation In

studies of autonomic activation, elevation in blood
pressure and cardiac
acceleration were measured in phobics during anticipation

and performance of
intimidating tasks corresponding to strong, medium, and

weak strength of per
ceived self-efficacy. Following the test for autonomic

reactions, subjects received
guided mastery experiences until they perceived themselves

to be maximally
self-efficacious on all of the previous coping activities.

Then their autonomic
reactions were again measured. Figure 1 shows the mean

change from the baseline level of heart rate and
blood pressure as a function of differential strength of

Subjects were viscerally unperturbed by coping tasks they

regarded with utmost
5 4 3 2 H E A R T R A T E A N T I C I P A T O R Y P E R F O
R M A N C E ~ 2 8 1 S M W 8 2 S I S ) S ( M 1 S ( W )
8 7 6 5 3 2 0 1 S Y S T O L I C 8 1 S M W 8 2 S ( S ) 5 ( M
) S ( W ) 6 5 D I A S T O L I C 8 1 S M W 8 2 5 ( 5 ) 5 ( M
) S ( W ) S T R E N G T H O F S E L F P E R C E P T S O F E
F F I C A C Y F i g u r e 1 M e a n c h a n g e f r o m t h
e b a s e l i n e l e v e l i n h e a r t r a t e a n d b l
o o d p r e s s u r e d u r i n g a n t i c i p a t o r y a
n d p e r f o r m a n c e p e r i o d s a s a f u n c t i o
n o f d i f f e r e n t i a l s t r e n g t h o f p e r c e
i v e d s e l f e f f i c a c y . B r e f e r s t o b a s e
l i n e l e v e l , a n d S , M , a n d W s i g n i f y s t
r o n g , m e d i u m , a n d w e a k s t r e n g t h s o f
p e r c e i v e d s e l f e f f i c a c y , r e s p e c t i
v e l y . F o r e a c h p h y s i o l o g i c a l m e a s u
r e t h e f i g u r e o n t h e l e f t i n e a c h p a n e
l s h o w s t h e a u t o n o m i c r e a c t i o n s r e l
a t e d t o s e l f e f f i c a c y b e l i e f s o f d i f
f e r i n g s t r e n g t h s ( p e r f o r m a n c e a r o
u s a l a t p e r c e i v e d w e a k s e l f e f f i c a c
y i s b a s e d o n o n l y a f e w s u b j e c t s w h o w
e r e a b l e t o e x e c u t e o n l y p a r t i a l p e r
f o r m a n c e s ) . T h e f i g u r e o n t h e r i g h t
o f t h e s a m e p a n e l s h o w s t h e a u t o n o m i
c r e a c t i o n s t o t h e s a m e s e t o f t a s k s a
f t e r s e l f b e l i e f s o f e f f i c a c y w e r e s
t r e n g t h e n e d t o t h e m a x i m a l l e v e l ( B
a n d u r a , R e e s e , & A d a m s , 1 9 8 2 ) . MEAN
CHANGE IN HEART RATE MEAN CHANGE IN BLOOD PRESSURE
MEAN CHANGE IN BLOOD PRESSURE 2 5 2 3 2 5 2 3 2 5 2 3 2
5 2 0 0
self-efficacy. However, on tasks about which they had

moderate doubts about
their coping efficacy, their heart rate accelerated and

their blood pressure rose
during anticipation and performance of the activities. When

presented with tasks
in their weak self-efficacy range, most subjects promptly

rejected them as too far
beyond their coping capabilities to even attempt. Indeed,

only a few subjects
were able to do any of them. Although lack of coping action

precluded a mean
ingful analysis of performance arousal, data from the

anticipatory phase shed
light on how autonomic reactions change when people

withdraw from transac
tions with threats they judge will overwhelm their coping

capabilities. Cardiac
reactivity promptly declined but blood pressure continued

to climb. After per
ceived self-efficacy was strengthened to the maximal level,

everyone performed
these previously intimidating tasks without any autonomic

activation. Heart rate is affected more quickly than blood
pressure by personal restruc
turing of intimidating task demands, which may explain the

different pattern of
autonomic reactivity at the extreme level of perceived

self-inefficacy. Catechol
amines, which govern autonomic activity, are released in

different temporal pat
terns during encounters with external stressors (Mefford et

aI., 198 I). Heart rate
is especially sensitive to momentary changes in

catecholamine patterns, with epi
nephrine, which is rapidly released, having a more

pronounced effect on cardiac
activity than on arterial pressure. Catecholamine

Activation Investigation of the biochemical effects of
perceived coping efficacy was
further extended by linking strength of perceived

self-efficacy to plasma cat
echolamine secretion (Bandura, Taylor, Williams, Mefford, &

Barchas, 1985).
The range of perceived coping efficacy in severe phobics

was broadened by
modeling which conveyed predictive information about the

phobic threat and
demonstrated effective ways of exercising control over it.

They were then pre
sented with coping tasks they had previously judged to be

in their low, medium,
and high self-efficacy range, during which continuous blood

samples were
obtained through a catheter. Figure 2 presents graphically

the microrelation between self-efficacy beliefs
and plasma catecholamine secretion. Levels of epinephrine,

norepinephrine, and
dopac, a dopamine metabolite, were low when phobics coped

with tasks in their
strong efficacy range. Self-doubts in coping efficacy

produced substantial in
creases in these catecholamines. When presented with tasks

that exceeded their
perceived coping capabilities the phobics instantly

rejected them. Catechola
mines dropped sharply. The dopac response differs markedly

from the other catecholamines. Where
as epinephrine and norepinephrine dropped upon rejection

of the threatening task,
dopac rose to its highest level, even though the phobics

had no intention of cop
ing with the task. Dopac seems to be triggered by the mere

apperception of envi
ronmental demands overwhelming one's perceived coping

capabilities. These F i g u r e 2 M i c r o r e l a t i o
n b e t w e e n w e a k , m e d i u m , a n d s t r o n g s
e l f p e r c e p t s o f c o p i n g e f f i c a c y a n d
l e v e l o f p l a s m a c a t e c h o l a m i n e s e c r
e t i o n . T h e c o p i n g t a s k c o r r e s p o n d i
n g t o t h e w e a k s e n s e o f e f f i c a c y w a s r
e j e c t e d ( B a n d u r a , T a y l o r , W i l l i a m
s , M e f f o r d , & B a r c h a s , 1 9 8 5 ) . . 1 4 5 ~
. 1 3 0 ~ . 1 1 5 ~ . 1 0 0 S T R O N G M E D I U M W E A K
. 3 4 0 ~ . 3 2 0 I w . 3 0 0 0 . 2 8 0 ~ . 2 6 0 S T R O N
G M E D I U M W E A K . 6 7 5 ~ . 6 2 5 ~ . 5 7 5 . 5 2 5 S
T R O N G M E D I U M S T R E N G T H O F P E R C E I V E D
S E L F E F F I C A C Y W E A K J'vdoa N\fla31N
3NlilHd3Nld3~ON N\fld31N 3NI~iHd3Nld3 N\fla31N
data suggest that under some conditions plasma dopac could
reflect activity of
brain dopamine neurons. Such a central contribution would

be consistent with the
enhanced dopac concentrations observed with perceived

inefficacy to cope with a
task, as shown in Figure 2. After perceived coping

efficacy was strengthened to the maximal level by
guided mastery, performance of the previously intimidating

tasks no longer elic
ited differential catecholamine reactivity. Thus, the

elevated catecholamine se
cretions observed in the initial test resulted from a

perceived mismatch between
coping capabilities and task demands, rather than from

properties inherent in the
tasks themselves. The evidence across a variety of

biological indices is consis
tent in showing that biological stress reactions to coping

activities differ when
perceived self-efficacy differs, but biological reactions

are the same when
perceived self-efficacy is raised to the maximal level.

The crucial role of controllability in biological
activation is further shown in
microanalysis of changes in catecholamine secretion as

phobics gain mastery
over phobic threats through guided mastery treatment

(Bandura et aI., 1985). In
this approach, phobics quickly gain mastery over threats

through aided guided
mastery. As treatment progresses, the mastery aids are

discontinued to verify
that coping attainments stem from the exercise of personal
efficacy rather than
from mastery aids. Self-directed mastery experiences are

then arranged to
strengthen and generalize the sense of coping efficacy.

Figure 3 presents the
plasma catecholamine levels at five demarcated stages in

treatment. During the
initial phases of treatment, when phobics lacked a sense of

coping efficacy, even
the mere sight or minimal contact with phobic threat

activated catecholamine re
sponses. After participants gained controlling efficacy,

their catecholamine level
dropped and remained relatively low during the most

intimidating interactions
with the phobic threat. When they were asked to relinquish

all control, which left
them completely vulnerable, catecholamine reactivity

promptly rose. This pat
tern of results is in accord with a mechanism involving

controllability rather than
simple extinction or adaptation over time. Autonomic

arousal to stressors is reduced by self-knowledge that one
can
wield control over them at any time even though that

controlling capability is un
exercised (Glass, Reim, & Singer, 1971). Choosing not to

exercise control at a
particular time, but being able to do so whenever one

wants to, should be distin
guished from relinquished control in which one is deprived

of all means of con
trol while subjected to stressors. Relinquished control

leaves one completely vul
nerable, whereas freely usable control, though unexercised

at a particular
occasion, leaves one in full command. Opioid Activation

Endogenous opioids play a paramount role in the regulation
of pain
(Fanselow, 1986) and as mediators of the effects of

uncontrollable stressors on
immunocompetence (Shavit & Martin, 1987). Studies with

animals subjected to
painful stimulation show that stress can activate

endogenous opioids that block
pain transmission (Fanselow, 1986). Opioid involvement is

indicated by evi
dence that stress-induced analgesia is reduced by opiate

antagonists, is blocked
by adrenalectomy, and is reinstated by administering

corticosterone to adrenalec
tomized animals (Grau, Hyson, Maier, Madden, & Barchas,

1981; MacLennan et
al., 1982). It is not the physically painful stimulation,

per se, but the psychologi
cal stress over its uncontrollability that seems to be a

key factor in opioid activa
tion (Maier, 1986). Animals who can turn off shock

stimulation show no opioid
activation, whereas yoked animals who experience the same

shock stimulation
without being able to control its offset give evidence of

stress-activated opioids. Another line of research on
biological mediators examined the impact of per
ceived coping efficacy on endogenous opioid activation

(Bandura, Cioffi, Taylor,
Brouillard, 1988). Differential levels of perceived

cognitive self-efficacy were
induced by having subjects exercise control over the pace

of cognitive demands
or by having the same demands controlled externally at a

pace that strained cog
nitive capabilities. Strong perceived self-efficacy was

accompanied by low stress,
whereas subjects who judged themselves inefficacious to

cope with the cognitive
demands experienced high subjective stress and autonomic

arousal. Stress can
activate endogenous opioids that block pain transmission

(Kelley, 1986). Opioid
involvement is indicated by evidence that stress-induced

analgesia is reduced by
opiate antagonists, such as naloxone, which blocks opiate

receptors. To test for
opioid activation, subjects were therefore administered

either naloxone or an inert
saline solution, whereupon their pain tolerance was

measured at periodic inter
vals by cold pressor tests. Efficacious subjects, whose

high sense of control kept
stress low, gave no evidence of opioid activation in that

their pain tolerance was
unaffected by naloxone (Figure 3). In contrast, the

perceived self-inefficacious
subjects, who experienced high stress, gave evidence of

opiate-mediated analge
sia. They displayed high pain tolerance under saline, but

found pain difficult to
bear under naloxone opioid blockage. The greater the

decline in perceived
cognitive efficacy, the greater was the opioid activation.
Opioid and Cognitive Mechanisms in Pain Control Pain is a
complex psychobiological phenomenon, influenced by
psychosocial
factors, rather than simply a sensory experience arising

directly from stimulation
of pain receptors. The same intensity of pain stimulation

can give rise to differ
ent levels of conscious pain depending on how attention is

deployed, how the ex
perience is cognitively appraised, the coping strategies

used to modulate pain,
and on modeled reactions to nociceptive stimulation

(Cioffi, 1991; Craig, 1986;
Turk, Meichenbaum, & Genest, 1983). Pain can be regulated

through different
mechanisms. We have already examined how pain sensations

can be counteract
ed at the locus of pain receptors by opioid blockage. Pain

can also be regulated
by central processes involving attentional and other

cognitive activities that
reduce consciousness of pain sensations.
Figure 3 Percent change in pain tolerance as a function of

perceived self-efficacy to exercise control over cognitive
demands and whether people received saline or the opiate
antagonist, naloxone (Bandura, Cioffi, Taylor, &
Brouillard, 1988). There are several ways by which
perceived coping efficacy can bring relief
from pain by cognitive means. People who believe they can

alleviate pain enlist
whatever ameliorative skills they have learned and will

persevere in their efforts
to reduce the level of experienced pain. Those who judge

themselves as ineffica
cious give up readily in the absence of quick relief.
Consciousness has a very
limited capacity (Kahneman, 1973). It is hard to keep more

than one thing in
mind at the same time. If pain sensations are supplanted in

consciousness, they
are felt less. Dwelling on pain sensations only makes them

more noticeable and,
thus, more difficult to bear. Perceived self-efficacy can

lessen the extent to
which painful stimulation is experienced as conscious pain

by diverting attention
from pain sensations to competing engrossments. Thus, for

example, attentional
diversion enables long-distance runners to press on even

though their body is
wracked in pain. Were they to focus on their mounting pain

sensations they
could not continue for long. During deep engrossment in

activities people can be
come oblivious to chronic pain sensations. Finally, people

who believe they can
exercise some pain control are likely to interpret

unpleasant bodily sensations and
states more benignly than those who believe there is

nothing they can do to alle
viate pain (Cioffi, 1991). Construals that highlight the

sensory rather than the
aversive aspects of pain reduce distress and raise pain

tolerance (Ahles,
Blanchard, & Leventhal, 1983). Results of several lines

of research indicate that perceived self-efficacy can
mediate the analgesic potency of different psychological

procedures. Reese
(1983) found that cognitive pain control techniques,
self-relaxation, and placebos 40 HIGH SELF-EFFICACY
NALOXONE SALINE 30 20 10 0 -10 -20 0 5 LOW
SELF-EFFICACY NALOXONE SALINE 15 30 0 5 15 30 TIME OF
POST-INJECTION TESTS (min) % C H A N G E I N P A I N I D L
E R A N C E ( s e c )
all increase perceived self-efficacy to cope with and

ameliorate pain. The more
self-efficacious the people judged themselves to be, the

less pain they experienc
ed in later cold pressor tests, and the higher was their

pain threshold and pain tol
erance. Arbitrary persuasory influences, in the fonn of

bogus feedback that one's
pain tolerance is high or low compared to that of others,

similarly alters people's
beliefs in their efficacy to manage pain which, in turn,

raises and lowers their
actual pain tolerance, respectively (Litt, 1988). Change

in self-efficacy belief is a
better predictor of pain tolerance than is past level of

pain tolerance. Arbitrarily
instilled perceived inefficacy restricts pain coping

behavior even when the oppor
tunity to exercise personal control exists, whereas

heightened perceived self
efficacy largely overrides ostensible external constraints

on personal control of
pain. Biofeedback is widely used as a muscle relaxation

procedure to ameliorate
pain. Holroyd and his colleagues have shown that the

benefits of biofeedback
training stem more from boosts in perceived coping

efficacy than from the
muscular exercises themselves (Holroyd et al., 1984).
Perceived self-efficacy,
created by false feedback that one is a skilled relaxer

for controlling pain, pre
dicted reduction in tension headaches, whereas the actual

amount of change in
muscle activity achieved in treatment was unrelated to the

incidence of subse
quent headaches. Studies of alternative cognitive

mechanisms of pain tolerance
reveal that perceived self-efficacy to manage pain

predicts pain tolerance, where
as outcome expectations of the amount of pain anticipated

for engaging in aver
sive activities do not independently affect how much pain

people endure when
variations in perceived self-efficacy were controlled

(Williams, Kinney, & Falbo,
1989). That perceived self-efficacy makes pain easier to

manage is further cor
roborated by studies of acute and chronic clinical pain

(Council, Ahern, Follick
& Kline, 1988; Dolce, 1987; Lorig, Chastain, Ung, Shoor, &
Holman, 1989;
Manning & Wright, 1983; O'Leary, Shoor, Lorig, & Holman,

1988). At first sight, helplessness theory and
self-efficacy theory appear to be at
odds on how controlling efficacy relates to pain tolerance

and the mechanisms
mediating it. Endurance of pain is associated with

deficient control over stressors
in helplessness theory, but with controlling efficacy in

self-efficacy theory. There
are several possible explanations for this seeming

contradiction. It might be rea
soned, from research on stress-induced analgesia, that

coping efficacy may en
hance pain control mainly through nonopioid mechanisms.

Because a high sense
of coping efficacy renders aversive situations less

stressful, it would reduce stress
activation of opioids. Although there may be less opioid

blockage of pain, exer
cise of personal efficacy that occupies consciousness with

engrossing matters can
block awareness of pain sensations by a nonopioid

cognitive mechanism. A second plausible explanation for
the paradoxical findings is in tenns of the
markedly different consequences of control in the types of

coping situations used.
The exercise of control produces fundamentally different

conditions of pain stim
ulation in the situations commonly used in animal and human

studies of pain that
would argue for some opioid involvement with high

self-efficacy. In the usual
animal experimentation, behavioral control promptly

terminates pain stimulation.
By contrast, in the human situation, efficacious exercise

of cognitive control over
pain sensations enables people to tolerate high levels of

pain stimulation but, in
so doing, it promotes even more active engagement in

activities that can heighten
the level and duration of pain stimulation. A strong sense

of coping efficacy
often increases engagement in pain-generating activities to

the point where it can
create stressful predicaments. Thus, for example,
self-efficacious people suffer
ing from arthritis generate pain and discomfort when they

frrst take on more vig
orous activities. Activity eventually improves function and

reduces pain but in
the short term it increases pain and distress. Similarly,

people experience mount
ing pain the longer they keep their hand immersed in icy
water in the cold pressor
task. Indeed, in the latter situation, continued exercise

of controlling efficacy
through cognitive means eventually heightens pain to the

point where it begins to
overwhelm people's coping capabilities and they begin to

experience the intense
pain stimulation as unbearable. The stress of failing

control with mounting pain
in later stages of coping would activate opioid systems.

In this conception of the human coping process, both opioid
and nonopioid
mechanisms operate in the regulation of pain, but their

relative contribution
varies with degree of controlling efficacy and phases of

coping. A nonopioid
mechanism would subserve pain tolerance while cognitive

control effectively
shuts out pain sensations from consciousness or renders

then less aversive by
benign construal. But an opioid mechanism would come into

play in later phases
of coping when control techniques become insufficient to

attenuate mounting
pain or to block it from consciousness. Thus, opioid
activation would remain
low during successful phases, but high during the more

stressful failing phases of
cognitive control. Research in which exercise of personal

efficacy lowers stress
rather than fosters activities of mounting aversiveness

that eventually
overwhelms coping capabilities yields findings similar to

those from studies of
uncontrollable physical stressors (Bandura et al., 1988).

Perceived self
inefficacy raises pain tolerance through opioid activation.

Evidence for pain control through the dual mechanisms is
provided by a
study in which individuals were either taught cognitive

methods of pain control,
administered a placebo presented as a medicinal analgesic,

or they received no
intervention (Bandura, O'Leary, Taylor, Gauthier, &

Gossard, 1987). Following
the treatment phase, their perceived efficacy to control

pain, and to reduce it, and
their tolerance of cold pressor pain were measured.

Participants in all conditions
were then administered either naloxone, an opiate

antagonist, or an inert saline
solution, and thereafter their pain tolerance was measured

at periodic intervals. Training in cognitive control
heightened perceived self-efficacy to endure
and reduce pain (Figure 4). Placebo medication had a

differential impact on per
ceived efficacy to endure pain and perceived efficacy to

reduce its intensity. Peo
ple believed they were better able to withstand pain with
the aid of a supposedly
pain-relieving medication. However, success in reducing

experienced pain
depends on effective exercise of pain ameliorating skills,

which medication alone
does not provide. Placebo medication did not persuade

people that they became
more capable of reducing the intensity of pain. These

findings underscore the
value of measuring different aspects of perceived

self-efficacy in research
designed to elucidate the exercise of control over pain.

Perceived self-efficacy
predicted how well people managed pain. The stronger their

beliefs in their
ability to withstand pain, the longer they endured

mounting pain, regardless of
whether their perceived self-efficacy was enhanced by

cognitive means or by
placebo medication or varied preexistantly without any

intervention. A strong
sense of efficacy to endure pain predicts tolerance of

mounting pain when initial
differences in pain tolerance are controlled.
Figure 4 Percent change from pretest level in perceived

self-efficacy and pain tolerance achieved by people who
were taught cognitive pain control techniques,
administered a placebo, or received no intervention
(Bandura, O'Leary, Taylor, Gauthier, & Gossard, 1987).
The findings provide evidence for both an opioid-mediated
component and a
nonopioid component for attenuating the impact of pain

stimulation by cognitive
means. As can be seen in Figure 4, cognitive copers who

were administered
saline displayed a sizable increase in pain tolerance. In

contrast, when pain
reducing opioids are blocked by naloxone, cognitive copers

found it more
difficult to manage pain. However, cognitive copers were

able to increase their
pain tolerance even under opioid blockage, which lends

support for a nonopioid
component as well in the exercise of cognitive control.

For cognitive copers
administered saline the combined action of both opioid and

cognitive control
contributed to their ability to achieve a sizable increase

in pain tolerance.
366
0 /
C H
A N
I N
A I
L E
R
A N
( s e
c ) 50 NALOXONE SAL INE 40 30 20 10 0 CONTROL

PLACEBO COGNITIVE CONTROL -10 0 5 20 60 0 5 20 60 0
5 20 60 TIME OF POST-INJECTION TESTS (min) The
correlational findings shed some light on how different
forms of self
efficacy relate to opioid activation under different modes

of coping. Coping with
heightened pain under opioid blockage requires active

exercise of strategies for
alleviating pain rather than mere forbearance. People who

judge themselves to
be good pain copers would be especially distressed by their

eventual ineffective
ness to manage their pain. Thus, the degree of opioid

activation is best predicted
by perceived capability to reduce pain. The stronger the

subjects' perceived self
efficacy to reduce pain, the greater was the opioid

activation. The strength of this
relationship is further increased when initial ability to

tolerate pain is controlled
by partial correlation. The findings also provide some

evidence that placebo medication may acti
vate some opioid involvement. After the full time had

elapsed for naloxone to
exert its antagonistic effect, people in the naloxone

condition were less able to
tolerate pain than those who had been given saline. These
findings are in accord
with those of Levine and his associates showing that

endogenous opioids can be
activated by placebo medication to reduce postoperative

dental pain (Levine,
Gordon, & Fields, 1978; Levine, Gordon, Jones, & Fields,

1978). A socially ad
ministered placebo produces analgesia, whereas unsignaled

mechanical infusion
of the placebo that goes undetected by patients has no

analgesic effect (Levine &
Gordon, 1984). Placebo-induced analgesia may involve both a

nonopioid cogni
tive component and a stress analgetic component that is

antagonizable by
naloxone (Gracely, Dubner, Wolskee & Deeter, 1983). In the

study under
discussion, placebo medication had its major impact on

withstand pain. Therefore, it was this expression of

efficacy that predicted de
gree of opioid involvement. The strength of the placebo

response is predictable
from how the placebo affects perceived self-efficacy to

endure pain (Bandura et
al., 1987). People who judge themselves efficacious to

withstand pain given a
supposed medicinal aid are good pain endurers, whereas

those who continue to
distrust their efficacy to manage pain despite receiving

the placebo medication
are less able to bear pain. For people who lack assurance
in their efficacy, the
evident failure to achieve relief from pain, even with the
help of a medicinal
analgesic, is only further testimony for their coping

inefficacy. So far the discussion has focused on
self-efficacy regulation of pain through
cognitive supplanting or construal of pain sensations and

stress-induced opioid
activation. There is a third possible mechanism that merits

serious consideration
as well. Self-efficacy expectations may directly activate

the central nervous sys
tem to release pain-blocking opioids independently of

stress. Animals can learn
to activate their endogenous opioid systems anticipatorily

in the presence of cues
formerly predictive of painful experiences (Watkins &

Mayer, 1982). Such find
ings add some credence to the possibility of direct central

activation of opioid
systems. Perceived Coping Self-Efficacy and

Immunocompetence Several lines of evidence suggest that
psychosocial factors modulate the im
mune system in ways that can influence susceptibility to

illness (Coe & Levine,
1991; Kiecolt-Glaser & Glaser, 1987; Locke et al., 1985;

O'Leary, 1990). The
types of biological reactions that have been shown to

accompany perceived cop
ing efficacy, such as autonomic activation, catecholamines,

and endogenous
opioids, are involved in the regulation of the immune

system. There are three
major pathways through which perceived self-efficacy can

affect immune func
tion. They include mediation through stress, depression and
expectancy learning. Stress mediation. The ability to
exercise control over potential stressors can
have significant impact on the cellular components of the

immune system. Expo
sure to intermittent stressors without the ability to

control them causes impair
ment in various facets of immune function, whereas exposure

to the same stress
ful events, but with efficacy to control them, has no

adverse effects on immune
function (Coe & Levine, 1991; Maier et al., 1985). There is

evidence that some
of the immunosuppressive effects of inefficacy in

controlling stressors, such as
reduced natural killer cell cytotoxicity, are mediated by

release of endogenous
opioids (Shavit & Martin, 1987). When opioid mechanisms are

blocked by
opiate antagonists, the stress of coping inefficacy loses

its immunosuppressive
power. These findings are based on experimentation with

animals involving
uncontrollable physical stressors. Human coping involves

an important feature that is rarely examined in either
animal laboratory paradigms or human field studies. In

animal experimentation,
controllability is usually studied as a fixed dichotomous

property in which ani
mals either exercise complete control over physical

stressors or they have no
control whatsoever. In contrast, human coping usually

entails an ongoing pro
cess of developing one's coping efficacy rather than
unalterable self-inefficacy in
the face of unremitting bombardment by stressors. Most

human stress is activat
ed in the course of learning how to exercise control over

environmental demands
and while developing and expanding competencies. Stress

activated in the pro
cess of acquiring coping self-efficacy may have very

different effects than stress
experienced in aversive situations with no prospect in

sight of ever gaining any
self-protective efficacy. There are substantial

evolutionary benefits to experienc
ing enhanced immunocompetence during development of coping

capabilities
vital for effective adaptation. It would not be

evolutionarily advantageous if
acute stressors invariably impaired immune function,

because of their prevalence
in everyday life. If this were the case, people would

experience high
vulnerability to infective agents that would quickly do

them in. Efforts to determine the immunologic effects of
psychological stressors in
humans have relied extensively on correlational or

quasi-experimental studies in
which occurrences of life stressors are related to the

incidence of infectious ill
nesses or to indices of immunologic functioning (Jemmott &

Locke, 1984;
O'Leary, 1990; Palmblad, 1981). Exposure to stressors is

usually accompanied
by impainnent of the immune system. These lines of research

have clarified
some aspects of inefficacious control of stressors, but

experimental studies are
needed to verify the direction of causality. That stress

aroused while gaining coping mastery over acute stressors
can
enhance different components of the immune system is

revealed in a study of ex
posure to a chronic stressor with experimentally varied

perceived coping self
efficacy (Wiedenfeld et a1., 1990). In this experiment, we

measured in severe
phobics their strength of perceived coping self-efficacy,

autonomic and neuroen
docrine activation and different aspects of their immune

system at three
Figure S Changes in components of the immune system

experienced as percent of baseline values during exposure
to the phobic stressor while acquiring perceived coping
self-efficacy (Efficacy Growth) and after perceived coping
self-efficacy had been developed to the maximal level
(Maximal Efficacy). The baseline mean values for the
different immune functions were as follows: Total
lymphocytes = 1572; T lymphocytes = 1124; Helper T-cells =
721; Suppressor T-cells = 370; Helper/Suppressor ratio =
2.22; HLD-DR = 216 (Wiedenfeld, O'Leary, Bandura, Brown,
Levine, & Raska, 1990). 25 % C H A N G E F R O M B A S E
L I N E 20 15 10 LYMPHOCYTES TOTAL T CELLS HELPER T
CELLS 25 8 40 20 6 30 15 4 20 10 2 10 0 0 0
SUPPRESSOR T CELLS HELPER/SUPPRESSOR HLA·DR 8 EFFICACY
MAXIMAL 8 EFFICACY MAXIMAL 8 EFFICACY MAXIMAL GROWTH
EFFICACY GROWTH EFFICACY GROWTH EFFICACY 0 % C H A N G
E F R O M B A S E L I N E
phases--during a baseline control phase involving no

exposure to the phobic
stressor; a perceived self-efficacy acquisition phase in

which they gained an
increasing sense of coping efficacy over the stressor

through guided mastery; and
a perceived maximal self-efficacy phase during which they
coped with the same
stressor after they had developed a complete sense of

coping efficacy. As may be seen in Figure 5, development
of strong perceived self-efficacy to
control stressors had an immunoenhancing effect. However,

a small subgroup of
individuals exhibited a decrease in immune system status

during the efficacy
acquisition phase. The rate of efficacy acquisition is a

good predictor of whether
exposure to acute stressors enhances or attenuates the

immune system. Rapid
growth of perceived coping efficacy reduces stress with

concomitant immunoen
hancing effects, whereas slow growth of perceived coping

efficacy is associated
with prolonged high stress and immunosuppressing effects.

High autonomic and
neuroendocrine activation also attenuate components of the

immune system, but
their impact is somewhat weaker. Acquisition of

perceived self-efficacy to control stressors produced more
than simply transient changes in immunity. The increase in

immunologic compe
tence was generally sustained over time as evident in the

significantly higher
immune system status in the maximal perceived self-efficacy

phase than in the
baseline phase. Rapid growth of perceived self-efficacy

also predicted mainte
nance of immunocompetence during the maximal perceived

self-efficacy phase.
These findings indicate that vigorous mastery of chronic
stressors not only in
stills a strong sense of self-efficacy but leaves lasting

changes that can serve as
protective factors against adverse immunologic effects of

psychological stressors. The field of health functioning
has been heavily preoccupied with the physio
logically debilitating effects of stressors. Self-efficacy

theory also acknowledges
the physiologically strengthening effects of mastery over

stressors. A growing
number of studies are providing empirical support for

physiological toughening
by successful coping (Dienstbier, 1989). Any comprehensive

theory of psycho
social modulation of health functioning must specify the

determinants and mech
anisms governing both debilitating and toughening effects

of coping with
stressors. Depression mediation. Bereavement and

depression have been shown to
reduce immune function and to heighten susceptibility to

disease (Ader & Cohen,
1985; Coe & Levine, 1991). Depressive states are,

therefore, often associated
with increased incidence of infectious disease,

development and spread of malig
nant neoplasms and accelerated rate of tumor cell growth.

These findings sug
gest that another possible path of influence of coping

inefficacy on immunocom
petence operates through the mediating effects of

depression. A sense of ineffi
cacy to fulfill desired goals that affect evaluation of

self-worth and to secure
things that bring satisfaction to one's life creates

depression (Bandura, 1988b;
Kanfer & Zeiss, 1983). Moreover, a low sense of cognitive

efficacy to tum off
perturbing ruminations contributes to the depth, duration

and recurrence of bouts
of depression (Kavanagh & Wilson, 1989). Supportive

relationships help to lessen the aversive impact of adverse
life
events that can give rise to depression. When the

self-doubts concern one's
social capabilities, they induce depression both directly

and indirectly by
curtailing social relationships that can provide

satisfactions and buffer the effects
of chronic daily stressors (Cutrona & Troutman, 1986;

Holahan & Holahan,
1987a, I 987b). O'Leary and her associates report findings

that are in accord
with depression mediation of immunity (O'Leary et al.,

1988). A low sense of
efficacy to exercise control over one's health functioning

was accompanied by
high levels of stress and depression, each of which was, in

tum, associated with
lowered functioning of several facets of the immune

system. Central mediation. The central nervous system can
exert regulatory influ
ence on immune function. Thus, a third possible path of

influence of perceived
self-efficacy is through central expectancy modulation of

immunologic reactivity.
Ader and Cohen (1981) have shown in animal experimentation

that immune
function is influenceable by expectancy learning. In

studies with humans, induc
ed expectations have been shown to affect physical

reactions to allergens and
antigens (Fry, Mason, & Pearson, 1964; Smith & McDaniel,

1983). Experiences involving successful or failed efforts
to manage environmental
demands produce cognitive changes in beliefs about personal

efficacy that have
significant physiological consequences when the

environmental stressor is no
longer present (Bandura et al., 1988). Thereafter, mere

thoughts about one's cop
ing efficacy lower autonomic activation in those whose

had been enhanced, but heighten autonomic reactions in

those whose sense of
coping efficacy was diminished. Such findings raise the

possibility that situa
tionally aroused self-expectations of coping efficacy may

produce anticipatory
immunosuppressive or immunoenhancing effects. PERCEIVED

SELF-EFFICACY AND ADOPTION OF HEALTH PRACTICES Lifestyle
habits can enhance or impair health status. This enables
people to
exercise some control over their vitality and quality of

health. Fries and Crapo
(1981) have marshalled a large body of evidence that the

upper limit of the
human life span is fixed biologically. Figure 6 shows the

mortality curves for a
society at different periods of time. People are now living

longer. Psychosocial
factors partly determine how much of the potential lifespan
is realized and the
quality of life that is lived. The ideal outcome can be

approximated by exercising
control over modifiable factors that slow the process of

aging and forestall the
development of chronic diseases and infmnities. The goal is

to enable people to
live their expanded lifespan productively with minimum

dysfunction, pain and
dependence. The physical problems of "old age" get

compressed into a short
period at the very end of the lifespan. One makes a rapid,

dignified exit when a
vital system finally gives out.
Figure 6 National survival curves in the United States for

different periods of time. In the ideal survival curve for
a society, people exercise control over modifiable aspects
of disease and aging so they live their expanded lifespan
with minimum dysfunction (Fries, 1989). The impact of
efficacy beliefs on the modifiable aspects of health and
aging
constitutes the second major level of research. Perceived

self-efficacy affects
every phase of personal change-whether people even consider

changing their
health habits; whether they can enlist the motivation and

perseverance needed to
succeed should they choose to do so; and how well they

maintain the changes
they have achieved. Initiation of Personal Change

People's beliefs that they can motivate themselves and
regulate their own
behavior plays a crucial role in whether they even consider

changing detrimental
health habits or pursuing rehabilitative activities (Beck &
Lund, 1981; Brod &
Hall, 1984). The perceived inefficacy barrier to preventive

health is all too famil
iar in people's resignation concerning health risks, such

as smoking or obesity,
over which they can exercise control. They see little point
to even trying if they
believe they do not have what it takes to succeed. If they

make an attempt, they
give up easily in the absence of quick results or setbacks.

Efforts to get people to adopt health practices that
prevent disease rely
heavily on persuasive communications in health education

campaigns. Meyero
witz and Chaiken (1987) examined four alternative

mechanisms through which
health communications could alter health habits-by

transmission of factual
information; fear arousal; change in risk perception; and

enhancement of per
ceived self-efficacy. They found that health communications

fostered adoption of
preventive health practices primarily by their effects on

perceived self-efficacy. 100 1980 1960 75 1940 (J) z
1920 :> :> '" => 50 11'1 .... Z U '" W IL 25 o
o 10 20 30 40 50 60 70 80 90 100 AGE 1980
Analyses of how community-wide media campaigns change

health habits simi
larly reveal that both the preexisting and induced level of

play an influential role in the adoption and social

diffusion of health practices
(Maibach, Aora, & Nass, 199); Slater, ) 989). The stronger

the preexisting per
ceived self-efficacy, and the more the media campaigns
enhance people's self
regulative efficacy, the more they are to adopt the

recommended practices
(Figure 7). To help people change health-impairing habits

clearly requires a shift
in emphasis from trying to scare people into health, to

empowering them with the
tools and self-beliefs of efficacy to exercise control over

their health habits.
Figure 7 Path analysis of the influence of perceived

self-efficacy on health habits in community-wide programs
to reduce risk of cardiovascular disease. The initial
numbers on the paths of influence are the significant path
coefficients for adoption of healthy eating patterns; the
numbers in parentheses are the path coefficients for
regular exercise (Maibach, Flora, & Nass, 1991).
Achievement of Personal Change Effective self-regulation
of health behavior is not achieved through an act of
will. Jt requires development of self-regulative skills. To

build people's sense of
controlling efficacy, they must develop skills on how to

influence their own
motivation and behavior. In such programs, they learn how

to monitor the
behavior they seek to change, how to set short-range

attainable subgoals to moti
vate and direct their efforts, and how to enlist incentives

and social supports to
sustain the effort needed to succeed (Bandura, 1986). Once

empowered with
skills and self-belief in their capabilities, people are

better able to adopt behaviors
that promote health and to eliminate those that impair it.

They benefit more from
treatments for physical disabilities and their
psychological well-being is less
adversely affected by chronic impairments. A growing body

of evidence reveals that the impact of different
therapeutic
interventions on health behavior is partly mediated through

their effects on per
ceived self-efficacy. The stronger the perceived

self-efficacy they instill, the
more likely are people to enlist the personal resources and

sustain the level of
effort needed to adopt and maintain health-promoting

behavior. This has been
shown in studies conducted in such diverse areas of health

as enhancement of PREEXISTING SELF-EFFICACY COMMUNITY
HEALTH CAMPAIGN .35(.32) .16 (.16) CHANGE IN
SELF-EFFICACY ADOPTION OF HEALTH HABITS . 3 5 ( . 3 2 )
pulmonary function in patients suffering from chronic

obstructive pulmonary
disease (Kaplan, Atkins, & Reinsch, 1984); recovery of

cardiovascular function
in postcoronary patients and activity level following

cardiac surgery (Gortner &
Jenkins, 1990; Taylor, Bandura, Ewart, Miller, & DeBusk,

1985); reduction in
pain and dysfunction in rheumatoid arthritis (Lorig,

Chastain, et al., 1989;
O'Leary et al., 1988); amelioration of tension headaches

(Holroyd et al., 1984);
control of labor and childbirth pain (Manning & Wright,

1983); management of
chronic low back, neck and leg pain and impairment (Council
et al., 1988); stress
reduction (Bandura, Reese, & Adams, 1982; Bandura et al,

1985); adjustment to
abortion (Major et al., 1990); weight reduction (Bernier &
Avard, 1986; Glynn &
Ruderman, 1986; Jeffrey et al., 1984; Stotiand, Zuroff, &

Roy, 1991); exercise of
control over bulimic behavior (Love, Ollendick, Johnson, &

Schlezinger, 1985;
Schneider, O'Leary, & Agras, 1987; Wilson, Rossiter,

Kleifield, & Lindholm,
1986); reduction of cholesterol through dietary means

(McCann, Follette, Driver,
Brief, & Knopp, 1988); adherence to prescribed remedial

activities (Ewart,
Stewart, Gillilan, & Kelemen, 1986; McCaul, Glasgow, &

Schafer, 1987); adop
tion and long-term adherence to a regular program of

physical exercise
(Desharnais, Bouillon, & Godin, 1986; McAuley, 1991; Sallis

et al., 1986; Sallis,
Pin ski, Grossman, Patterson, & Nader, 1988); maintenance

of diabetic self-care
(Crabtree, 1986); effective management of sexual coercions

and contraceptive
use to avoid unwanted pregnancies (Levinson, 1986); control

of sexual practices
that pose high risk for transmission of AIDS (McKusick,

Coates, Morin, Pollack,
& Hoff, 1990; O'Leary, Goodhart, Jemmott, &

Boccher-Lattimore, 1991); and
exercise of control over drug use (Gossop, Green,

Phillips, & Bradley, 1990),
alcohol abuse (Sitharthan & Kavanagh, 1991; Solomon &

Annis, 1990), and
cigarette smoking (Coletti, Supnick, & Payne, 1985;

DiClemente et al., 1991;
McIntyre, Lichtenstein, & Mermelstein, 1983). Maintenance

of Personal Change Habit changes are of little consequence
unless they endure. Maintenance of
habit change relies heavily on self-regulatory capabilities

and the functional
value of the behavior. Development of self-regulatory

capabilities requires instil
}jng a resilient sense of efficacy as well as imparting

skills. Experiences in exer
cising control over troublesome situations serve as

self-efficacy builders. This is
an important aspect of self-management because if people

are not fully convinced
of their personal efficacy they rapidly abandon the skills

they have been taught
when they fail to get quick results or suffer reverses.

Studies of behavior that is
amenable to change, but which is difficult to sustain over

an extended period,
show that perceived self-inefficacy increases vulnerability

to relapse (Bandura,
1991). Moreover, perceived efficacy predicts how

participants are likely to re
spond to subsequent relapse, should it occur. Those who

have a strong belief in
their efficacy tend to regard a slip as a temporary setback

and reinstate control. In
contrast, those who distrust their self-regulative

capabilities display a marked
decrease in perceived self-efficacy after a slip and

relapse completely. People's beliefs about their efficacy
can be altered in four principal ways
(Bandura, 1986). The most effective way of instilling a

strong sense of efficacy
is through mastery experiences. Successes build a robust

sense of efficacy. Fail
ures undermine it, especially if failures occur early in

the course of events. The
second method is through modeling. The models in people's

lives serve as
sources of competencies and motivation. People partly judge

their capabilities in
comparison with others. Seeing people similar to oneself

succeed by perseverant
effort, raises observers' beliefs about their own

capabilities. The failures of
others, instill self-doubts about one's own ability to

master similar tasks. Social
persuasion is the third mode of influence. Realistic

boosts in efficacy can lead
people to exert greater effort, which increases their

chances of success. People
also rely partly on their physiological state in judging

their capabilities. They
read their anxiety arousal and tension as signs of

vulnerability to dysfunction. In
activities involving strength and stamina, people interpret

their fatigue, aches,
and pains as indicants of physical inefficacy. The fourth

way of modifying effi
cacy beliefs is to reduce people's physiological

overreactions or change how they
interpret their somatic sensations. Each of these modes

for enhancing self-efficacy can be put to the service of
developing the resilient sense of perceived efficacy needed

to override difficul
ties that inevitably arise from time to time. With regard
to the enactive mastery
mode, a resilient sense of personal efficacy is built

through structured demon
stration trails in the exercise of control over

progressively more challenging
tasks. For example, as part of instruction in cognitive

control strategies, arthritic
patients were given efficacy demonstration trials in which

they performed select
ed pain-producing activities with, and without, cognitive

control and rated their
pain level (O'Leary et al., 1988). Explicit evidence that

they achieved substantial
reduction in experienced pain with cognitive control

provided the patients with
persuasive demonstrations that they could exercise some

control over pain by
enlisting cognitive strategies. Self-efficacy confirming

trials not only serve as
efficacy builders, but put to trial the value of the

techniques being taught. Modeling influences, in which
other patients demonstrate how to cope with
difficult situations, reinstate control should a setback

occur and show that success
usually requires tenacious effort, can further strengthen

Moreover, modeled perseverant success can alter the

diagnosticity of failure
experiences as partly reflecting difficult situational

predicaments rather than
solely inherent personal limitations. Difficulties and

setbacks prompt redoubling
of efforts rather than provoke self-discouraging doubts

about one's coping
capabilities. Persuasory influences that instill

self-beliefs conducive to optimal utiHzation
of skills can also contribute to staying power. As a

result, people who are per
suaded they have what it takes to succeed and are told that
the gains achieved in
treatment verify their capability are more successful in

sustaining their altered
health habits over a long time than those who undergo the
same treatment with
out the efficacy-enhancing component (Blittner, Goldberg, &

Merbaum, 1978).
Successful persuasory efficacy enhancers do more than

convey positive apprais
als. In addition to raising people's beliefs in their

capabilities, they structure situ
ations for them in ways that bring success and avoid

placing them prematurely in
situations where they are likely to fail. By maintaining an

efficacious attitude that
gains are attainable when clients are beset with

self-doubts, they can be helped to
sustain their coping efforts in the face of reverses and

discouraging obstacles. SelfEfficacy in the Causal
Structure of Social Cognitive Theory Perceived
self-efficacy operates within social cognitive theory as
one of
many determinants that regulate human motivation, emotional

activation and
behavior (Bandura, 1986). In addition to the regulative

function of self-efficacy
beliefs, outcome expectations concerning the effects that

may flow from different
forms of behavior contribute to health behavior. These
outcome expectations
take the form of detrimental or beneficial physical

effects, favorable or adverse
social consequences, and positive or negative

self-reactions. Cognized goals and
internal standards rooted in value systems also create

self-incentives and guides
for health behavior. Cognitive and behavioral strategies

aid in the translation of
self-beliefs, outcome expectations and personal goals to

successful action. Dzewaltowski and his colleagues tested
the predictiveness of a subset of
sociocognitive determinants that included perceived

self-efficacy to adhere to
health-promoting behavior, outcome expectations of physical

benefits, and self
evaluative reactions to one's behavioral attainments

(Dzewaltowski, 1989;
Dzewaltowski, Noble, & Shaw, 1990). These three factors

accounted for a
substantial amount of variance in health-promoting

behavior. A number of conceptual models, founded largely
on expectancy-value
theory, have been devised to predict adoption of preventive

health practices
(Becker & Maiman, 1983; Rosenstock, Strecher, & Becker,

1988; Schwarzer,
1992). They include such variables as the perceived

severity of, and susceptibil
ity to, a health threat, the perceived effectiveness and

costs of the protective
action, and the anticipated outcomes for different courses

of action. Recent
efforts to increase the predictiveness of such models have
added the self-efficacy
determinant to the usual set of predictors. Self-efficacy

beliefs make a significant
independent contribution to health behavior within the

expanded models (Ajzen
& Madden, 1986; De Vries, Dijkstra, & Kuhlman, 1988; McCaul

et al., 1987;
McCaul, O'Neill, & Glasgow, 1988; O'Leary et al., 1991;

Schifter & Ajzen,
1985; Schwarzer, 1992). The development of a comprehensive

conceptual model of health behavior
should be guided by the principle of parsimony. The extant

conceptual models
multiply predictive factors in two ways: through redundancy

by including essen
tially the same determinant under different names; and by

fractionating a higher
order construct into facets and labeling them with

dissimilar names as though
they represented fundamentally different classes of

determinants. These points
can be illustrated by comparing the causal structure of

social cognitive theory and
the Ajzen and Fishbein (1980) theory of reasoned action.

According to the theory of reasoned action, the intention
to engage in a
behavior is governed by attitudes toward the behavior and

by subjective norms.
Attitude is measured in terms of perceived outcomes and

the value placed on
those outcomes. Norms are measured by perceived social

pressures by signifi
cant others and one's motivation to comply with their

expectations. The latter
factor corresponds to expectations of social outcomes for

a given behavior. In
social cognitive theory normative influences regulate

actions through two regu
latory processes-social sanctions and instated

self-sanctions. Norms influence
behavior anticipatorily by the social consequences they

provide. Behavior that
fulfills social norms gains positive social reactions;

behavior that violates social
norms brings social censure. People do not act like

weathervanes, constantly
shifting their behavior to conform to whatever others

might want. Rather, they
adopt certain standards of behavior for themselves and

regulate their behavior
anticipatorily through self-evaluative consequences. They

do things that give
them self-satisfaction, they refrain from behaving in ways

that violate their stand
ards because it will bring self-censure. Social norms

convey behavioral stand
ards. Adoptions of modeled standards creates a

self-regulatory system that
operates through internalized self sanctions. If one looks

beyond the divergent terminology, the determinants singled
out
by Fishbein and Ajzen overlap with a subset of the

determinants encompassed by
social cognitive theory. Attitudes and subjective norms

represent different clas
ses of outcome expectations. Intentions are essentially

goal representations. The
overlap in determinants is confirmed empirically by
Dzewaltowski and his col
leagues (Dzewaltowski, 1989; Dzewaltowski et al., 1990).

After establishing the
contribution to health behavior of three sociocognitive

determinants (i.e., per
ceived self-efficacy, outcome expectations of physical

benefits, and self-evalu
ative consequences), the investigators added attitudes and

norms from the Fish
bein and Ajzen model to the multiple regression equation.

Attitudes and norms
did not account for any unique variance in health behavior

over and above the
social cognitive determinants. Many of the constructs of

different conceptual models of health behavior
measure similar classes of determinants. But the theories

differ in how well they
are grounded in knowledge of regulatory mechanisms and in

principles for con
structing effective interventions. Health belief models

seem to be concerned
mainly with predicting health behavior, but they say little

about how to design
programs to change it. Indeed, some of the determinants

tend to be conceptualiz
ed in ways that do not lend themselves easily to feasible

guidelines for personal
change. For example, consider the intention determinant in

the model of reason
ed action. Intentions presumably control actions. But one

is left with consid
erable prescriptive ambiguity on how to change attitudes

and social norms that
are said to create intentions. In contrast, research

conducted within the socio
cognitive framework has given us a large body of knowledge

on how to develop
self-regulatory capabilities, structure goals and feedback

systems, and mobilize
social supports to foster and maintain changes in health

practices (Bandura, 1986;
Holroyd & Creer, 1986; Maccoby & Solomon, 1981; Puska,

Nissinen, Salonen,
& Toumilehto, 1983; Winett, King, & Altman, 1989).

Self-Regulatory Model of Health Promotion and Risk
Reduction Hea1th care expenditures are soaring at a rapid
rate (Fuchs, 1990). Despite
the huge outlays for health services people are often

poorly served by traditional
health delivery systems. With people living longer and the

need for health care
services rising with age, societies are confronted with

major challenges on how to
keep people healthy throughout their lifespan, otherwise

they will be swamped
with burgeoning health costs. This requires intensifying

health promotion efforts
and restructuring health delivery systems to make them

more productive. Health promotion and risk reduction
programs are often structured in ways
that are costly, cumbersome and minimally effective. If

services are funneled
through physicians, it often creates a bottleneck in the

system. Many of them do
not know how to change high risk behavior. Even if they

did, they cannot spare
much time for any individual or make much money doing it.
The net result is
minimal prevention and costly remediation. Self-management

programs based on the self-efficacy model improve the
quality of health and greatly reduce utilization of

medica1 services. DeBusk and
his colleagues devised an efficacy-based model combining

self-regulatory princi
ples with computerized implementation that promotes habits

conducive to hea1th
and reduces those that impair it. This computerized

self-regulatory system
equips participants with the skills and personal efficacy

to exercise self-directed
change. It includes exercise programs to build

cardiovascular capacity; dietary
programs to reduce risk of heart disease and cancer;

weight reduction programs;
smoking cessation programs; and stress management programs

to reduce the
wear and tear on the body. For each risk factor,

individuals are provided with detailed guides on how to
alter their habits, along with a self-monitoring, goal

setting, and feedback system
to facilitate their efforts at self-directed change. A

single program implementor,
assisted by the computerized system, oversees the

behaviora1 changes of large
numbers of participants. Figure 8 portrays the structure

of the self-regulatory
system. At selected intervals, the computer generates and

mails to participants
individually-tailored guides for self-directed change that

specify subgoals and
portray graphica1ly the progress patients are making toward
their subgoals and
their month-to-month changes. Self-efficacy ratings

identify areas of vulnerabil
ity and difficulty. The participants, in tum, send data to

the implementor on the
changes they have achieved and their level of persona1

efficacy in the various
domains for the next cycle of self-directed change. The

program implementor
maintains telephone contact with the participants and is

available to provide them
extra guidance and support should they encounter

difficulties. The implementor
also serves as the liaison to medical personnel, who are

called upon when their
expertise is needed. SELF REGULATORY DELIVERY SYSTEM
Figure 8 Computer-assisted self-regulatory system for

altering health habits. Evidence for the effectiveness of
this self-regulatory system is available from
a cholesterol reduction program conducted with employees

with elevated choles
terol levels drawn from work sites. Each 1 % reduction in

serum cholesterol
achieves about a 2% reduction in risk of heart attack. The

program, which
sought to reduce consumption of cholesterol and saturated

fat, required a total of
two hours per employee. The participants lowered their

intake of saturated fat
and achieved significant reductions in serum cholesterol by

this means (Figure
9). They realized an even larger risk reduction if their

spouses took part in the
dietary change program as well. Some individuals have a

genetic metabolic defi
ciency for processing saturated fats and cholesterol so

their body produces high
levels of cholesterol even though they do not consume much

fat. Among patients
with elevated plasma cholesterol, the more room for dietary

reduction of saturat
ed fats the more substantial the reductions they achieve by

self-regulative means
in plasma cholesterol. The success of this system to

reduce morbidity and mor
tality in post-coronary patients is currently being

compared against the usual
medical post-coronary care. In this effort to reduce the

likelihood of future heart
attacks, a number of risk factors, including obesity,

elevated cholesterol,
smoking, sedentariness, and stress proneness, are selected

for change. The self-regulatory system is well received
by participants because it is indi
vidually tailored to their needs; it provides them with

continuing personalized
guidance and informative feedback that enables them to

exercise considerable
control over their own change; it is a home-based program

that does not require
any special facilities, equipment, or attendance at group

meetings that usually
have high drop-out rates; and it can serve large numbers

of people simulta
neously. The substantial productivity gains are achieved by

process innovations
combining self-regulatory and computer technologies that
provide effective
Self-Efficacy Mechanism PHYSICIAN PROGRESS REPORTS

PROGRAM IMPLEMENTOR PHONE CONTACT COMPUTERIZED SYSTEM
DATA BASE SELF -REGULATORY CHANGE PROGRAMS 379 PATIENT
A. Bandura
health-promoting services in ways that are individualized,

intensive, highly
convenient and inexpensive. 30 30 SELF-REGULATION

CONTROL =25 ~ 25 Cl g ...J 0 II: W 20 20 l(/)
W ...J 0 J: 0 < :E 15 15 (/) < ...J Q. ~ z
10 10 0 t= (,) :J 0 W II: 5 5 o SUBJECT SUBJECT
SMALL LARGE & SPOUSE ROOM FOR DIETARY CHANGE
Figure 9 Levels of reduction in plasma cholesterol achieved

with the computerized self-regulation system. The panel
on the left summarizes the mean cholesterol reductions
achieved in applications in the workplace by participants
who used the system either by themselves, along with their
spouses, or did not receive the system to provide a
control baseline. The right panel presents the mean
cholesterol reductions achieved with the selfregulative
system by patients whose daily cholesterol and fat intake
was high or relatively low at the outset of the program.
Self-Management of Chronic Diseases Chronic disease has
become the dominant fonn of illness and the major cause
of disability. Such diseases do not lend themselves well to

biomedical approaches
devised primarily to treat acute illness. The treatment of

chronic disease must
focus on self-management of physical conditions over time

rather than on cure.
This requires, among other things, pain amelioration,

enhancement and
maintenance of functioning with growing physical

disability and development of
self-regulative compensatory skills. Lorig and her

colleagues have devised a
prototypic model for the self-management of different

types of chronic diseases o
(Lorig, Seleznick, et aI., 1989). The self-management

skills include cognitive
pain control techniques, self-relaxation, proximal goal

setting to increase level of
activity and use of self-incentives as motivators, problem

solving and self
diagnostic skills for monitoring and interpreting one's

health status and skills in
locating community resources, managing medication programs,

and effective
ways of dealing with physicians and other aspects of health

care systems to
optimize health benefits. Participants are taught how to

exercise some control
over their physical condition through modeling of

self-management skills, guided
mastery practice, informative feedback and efficacy

demonstration trials. The effectiveness of this
self-regulative approach has been tested extensively
for ameliorating the debility and chronic pain of arthritis

(Lorig, Seleznick, et aI.,
1989). Patients suffering from rheumatoid arthritis,

substantially improve their
psychophysical functioning following treatment compared to

matched controls
who received an arthritis helpbook describing

self-management techniques and
were encouraged to be more active (O'Leary et aI., 1988).

The self-management
program increased patients perceived self-efficacy to

reduce pain and other debil
itating aspects of arthritis, and to pursue potentially

painful activities (Figure 10).
The treated patients reduced their pain and inflammation in
their joints, and were
less debilitated by their arthritic condition. The higher

their perceived coping ef
ficacy, the less pain they experienced, the less they were
disabled by their arthri
tis, and the greater the reduction they achieved in joint

impairment. The more
efficacious were also less depressed, less stressed, and

they slept better. The treatment did not alter immunologic
function, but significant relation
ships were found between perceived coping efficacy and

immunologic indices.
There is some evidence that in the arthritic disorder the

suppressor T-cell func
tion of the immune system is depressed. This results in

proliferation of anti
bodies, which is aided by helper T-cells. Arthritis is an

autoimmune disorder in
which the immune system produces antibodies that destroy

normal tissues of the
body. Increases in suppressor T-cells, which tend to

inhibit production of anti
bodies, suggest improvement in the immune system for this

disorder. Perceived
coping efficacy was associated with increases in the number

of suppressor T -cells
and with a decrease in the ratio of helper to suppressor

T-cells. In a follow-up assessment conducted four years
later, arthritis patients who
have had the benefit of self-management training displayed

increased self
efficacy, reduced pain, much lower utilization of medical

services (43%), and
slower biological progression of their disease over the
four-year period (Holman,
Mazonson, & Lorig, 1989; Lorig, 1990). These changes are

shown in Figure 11.
Enhancement of functioning despite some biological

progression of the disease
provides further testimony that functional limitations may

be governed more by
self-beliefs of capability than by degree of actual

physical impairment (Baron,
Dutil, Berkson, Lander, & Becker, 1987). Tests of

alternative mediating mecha
nisms reveal that neither increases in knowledge nor degree

of change in health
behaviors are appreciable predictors of health functioning

(Lorig, Chastain, et aI.,
1989; Lorig, Seleznick, et aI., 1989). However, both

baseline perceived selfF i g u r e 1 0 C h a n g e s e x h
i b i t e d b y a r t h r i t i c p a t i e n t s i n p e r
c e i v e d s e l f e f f i c a c y a n d r e d u c t i o n
i n p a i n a n d i m p a i r m e n t o f j o i n t s a s a
f u n c t i o n o f t r a i n i n g i n s e l f r e g u l a
t o r y t e c h n i q u e s ( O ' L e a r y , S h o O T , L
o r i g , & H o l m a n , 1 9 8 8 ) . NOI.L:)Nn.:l .LNIOr
NI .LN3V113AOtldVIII (%) A.L1l18VSIO NI .LN3V113AOtldVIII
(%) NIVd NI NOI.L:)n03t1 (%) A:)V:)I.:I.:I3-.:I13S .:10
HHJN3t1.LS NI 3EJNVH:) (%) 3 0 3 0 2 0 2 0 1 0 o 1 0 1 0 S
E S E S E P A I N F U N C T I O N A R T H R I T I S P E R C
E I V E D S E L F E F F I C A C Y 1 5 1 0 5 o A V E R A G E
H I G H E S T D I S A B I L I T Y P A I N P A I N 1 5 5 o T
r e a t m e n t C o n t r o l J O I N T F U N C T I O N
efficacy and changes in perceived self-efficacy to exercise

some control over
one's arthritic condition instilled by treatment explain

more than 40% of the
variance in pain four years later (Lorig, 1990). When

patients are equated for
degree of physical debility, those who believe they can
exercise some influence
over how much their arthritic condition affects them lead

more active lives and
experience less pain (Shoor & Holman, 1984). This

self-management model
lends itself well to other types of chronic diseases. +20

+10 0 W CJ Z c( J: ·10 0 ~ z w 0 II: ·20 w
II. ·30 .40 PHYSICIAN VISITS
Figure 11 Enduring healthful changes achieved by training

in self-management of arthiritis as revealed in a
follow-up assessment four years later. The 9% biological
progression of the disease is much less than the 20%
disease progression one would normally expect over four
years for this age group. Plotted from data of Lorig,
1990. IMPACT OF PROGNOSTIC JUDGMENTS ON SELF -EFFICACY
AND HEALTH OUTCOMES Much of the work in the health field
is concerned with diagnosing maladies,
forecasting the likely course of different physical

disorders and prescribing rem
edies. Medical prognostic judgments involve probabilistic

inferences from
knowledge of varying quality and inclusiveness about the

multiple factors gov
erning the course of a given disorder. Because psychosocial

factors account for
some of the variability in the course of health functioning

their inclusion in prog
nostic schemes will enhance their predictive power.

Prognostic judgments can
alter perceived self-efficacy in ways that affect health

outcomes rather than
simply serve as nonreactive forecastings of things to come

(Bandura, 1992). DISIEASE PROGRESSION SELF EFFICACY PAIN
A. Bandura Scope of Prognostic Schemes One important
issue regarding prognosis concerns the range of factors
included in a prognostic scheme. As previously noted,

level of health functioning
is determined not only by biologically-rooted factors but

also by patients' self
beliefs and a system of social influences that can enhance

or impede the progress
they make. If one takes no notice of psychosocial

determinants one is left with
puzzling variability in the courses that health changes

take and unexplained dif
ferential functional attainments of people who are equally

physically impaired.
Thus, arthritics with deformed hands may lead fulfilling

productive lives, where
as others with minimal arthritic impairment become

despondent and abandon
activities. Neither biochemical laboratory tests nor

measures of degree of actual
physical impairment predicts functional attainments (Baron

et al., 1987). Results
of a program of research on enhancement of perceived

self-efficacy for post
coronary recovery indicate that strength of perceived

self-efficacy is a psycho
logical prognostic indicator of the course that health

outcomes are likely to take. About half the patients who
experience myocardial infarctions have uncom
plicated ones (DeBusk, Kraemer, & Nash, 1983). Their heart

heals rapidly, and
they are physically capable of resuming an active life.

However, the psychologi
cal and physical recovery is slow for patients who believe

they have an impaired
heart. They avoid physical exertion. They fear that they

cannot handle the
strains in their vocational and social life. They give up
recreational activities.
They fear that sexual activities will do them in. The

recovery problems stem
more from patients' beliefs that their cardiac system has

been impaired than from
physical debility. The rehabilitative task is to convince

patients that they have a
sufficiently robust cardiovascular system to lead

productive lives. Psychological recovery from a heart
attack is a social, rather than solely an
individual matter. Virtually all of the patients are

males. The wives' judgments
of their husbands' physical and cardiac capabilities can

aid or retard the recovery
process. The direction that social support takes is partly

determined by percep
tions of efficacy. Spousal support is likely to be

expressed in curtailment of
activity if the husband's heart function is regarded as

impaired, but as encourage
ment of activity if his heart function is judged to be

robust. In the program de
signed to enhance postcoronary recovery (Taylor et aI.,

1985), the treadmill was
used to raise and strengthen spousal and patients' beliefs

in their cardiac
capabilities. Several weeks after patients have had a

heart attack we measured their beliefs
about how much strain their heart could withstand. They

then performed a
symptom-limited treadmill, mastering increasing workloads

with three levels of
spouse involvement in the treadmill activity. The wife was
either uninvolved in
the treadmill activity; she was present to observe her

husband's stamina as he
performed the treadmill under increasing workloads; or she

observed her
husband's performance, whereupon she performed the

treadmill exercises herself
to gain firsthand information of the physical stamina it

requires. We reasoned
that having the wives personally experience the

strenuousness of the task, and
seeing their husbands match or surpass them, should

convince them that their
husband has a robust heart. After the treadmill

activities, couples were fully informed by the cardiologist
about the patients' level of cardiac functioning and their

capacity to resume activ
ities in their daily life. If the treadmill is interpreted

as an isolated task, its impact
on perceived cardiac and physical capability may be

limited. In order to achieve a
generalized impact of enhanced self-efficacy on diverse

domains of functioning,
the stamina on the treadmill was presented as a generic

indicant of cardiovascular
capability-that the patients' level of exertion exceeded

whatever strain everyday
activities might place on their cardiac system. This would

encourage them to
resume activities in their everyday life that place weaker

demands on their
cardiac system than did the heavy workloads on the

treadmill. The patient's and
spouse's beliefs concerning his physical and cardiac
capabilities were measured
before and after the treadmill activity, and again after

the medical counseling.
Figure 12 Changes in perceived physical and cardiac

efficacy as a function of level of spouse involvement,
patients' treadmill exercises, and the combined influence
of treadmill exercises and medical counseling. Perceived
efficacy was measured before the treadmill (Pre), after
the treadmill (T), and after the medical counseling (C)
(Taylor, Bandura, Ewart, Miller, & DeBusk, 1985). 80 >70
u c( ~ 60 lL lL W • 50 lL ...J w 1fl40 o w >
w U 0: wao Q. lL 70 o J: 60 I~ 50 w 0: 40 IIfl
30 20 PRE T C PRE T C PRE T C PERCEIVED PHYSICAL AND
EMOTIONAL SELF-EFFICACY PRE SPOUSE ABSENT SPOUSE
OBSERVES SPOUSE PARTICIPATES T C PRE T C PRE T C
PERCEIVED CARDIAC SELF-EFFICACY SPOUSE SPOUSE A. Bandura
Figure 12 shows the patterns of changes in beliefs
concerning the patients'
physical and cardiac capabilities at different phases of

the experiment under dif
ferent levels of spousal involvement in the treadmill

activity. Treadmill perform
ances increased patients' beliefs in their physical and

cardiac capabilities. Initial
ly the beliefs of wives and their husbands were highly

discrepant-husbands
judged themselves moderately hearty, whereas wives judged

their husbands' car
diac capability as severely impaired and incapable of

withstanding physical and
emotional strain. Spouses who were either uninvolved in, or

merely observers of,
the treadmill activity, did not change their considerable

doubts about their hus
bands' physical and cardiac capabilities. Even the detailed

medical counseling by
the cardiology staff did not alter their preexisting
beliefs of cardiac debility.
However, wives who had personally experienced the

strenuousness of the tread
mill were persuaded that their husbands had a sufficiently

robust heart to with
stand the normal strains of everyday activities. The

participant experience appar
ently altered spousal cognitive processing of treadmill

information, giving greater
weight to indicants of cardiac robustness than to

symptomatic signs of cardiac
debility. Efficacy beliefs affect receptivity to

prognostic information. Thus, the
change in perceived efficacy made the wives more accepting

of the medical
counseling. Following the medical counseling, couples in

the participant spouse
condition had congruently high perceptions of the

patients' cardiac capabilities. The findings further show
that beliefs of cardiac capabilities can affect the
course of recovery from myocardial infarction. The higher

the patients' and the
wives' beliefs in the patients' cardiac capabilities, the

greater was the patients'
cardiovascular functioning as measured by peak heart rate

and maximal workload
achieved on the treadmill six months later. The joint

belief in the patients' car
diac efficacy proved to be the best predictor of cardiac

functional level. Initial
treadmill performance did not predict level of

cardiovascular functioning in the
follow-up assessment when perceived efficacy is partialled

out. But perceived
cardiac efficacy predicted level of cardiovascular

functioning when initial tread
mill performance was partialled out. Wives who believe

that their husbands have a robust heart are more likely to
encourage them to resume an active life than those who

believe their husband's
heart is impaired and vulnerable to further damage. The

positive relation be
tween the wife's perceptions of her husband's cardiac

capability and his treadmill
accomplishments months later is, in all likelihood, partly

mediated by spousal
encouragement of activities during the interim period.

Pursuit of an active life
improves the patient's physical capability to engage in

activities without over
taxing their cardiovascular system. Self-Validating

Potential of Prognostic Judgments Health outcomes are
related to predictive factors in complex, multidetermin
ed and probabilistic ways. Prognostic judgments, therefore,

involve some degree
of uncertainty. The predictiveness of a given prognostic

scheme will depend on
the number of relevant predictors it encompasses, the

relative validities and inter
relations of the predictors, and the adequacy with which

they are measured.
There is always leeway for expectancy effects to operate

because prognostic
schemes rarely include all of the relevant biological and

psychosocial predictors
and even the predictors that are singled out usually have
less than perfect valid
ity. Based on selected sources of information,
diagnosticians form expectations
about the probable course of a disease. The more confident

they are in the
validity of their prognostic scheme, the stronger are

their prognostic expectations. Prognostic judgments are
not simply nonreactive forecasts of a natural course
of a disease. Because prognostic information can affect

patients' beliefs in their
physical efficacy, diagnosticians not only foretell but may

partly influence the
course of recovery from disease. Prognostic expectations

are conveyed to pa
tients by attitude, word, or the type and level of care

provided them. Prognostic
judgments have a self-confmning potential. Expectations can

alter patients' self
beliefs and behavior in ways that confirm the original

expectations. Evidence in
dicates that the self-efficacy mechanism operates as one

important mediator of
self-confmning effects (Bandura, 1992; Litt, 1988).

Analysis of self-confirming
processes has focused mainly on how people's self-beliefs

of efficacy and behav
ior are affected by what they are told about their

capabilities. Other evidence
suggests that prognostic judgments may bias how people are

treated as well as
what they are told. Individuals treat others differently

under high than under low
expectations in ways that tend to confirm the original

expectations (Jones, 1977;
Jussim, t 986). Under induced high expectations individuals
generally pay more
attention to those in their charge, provide them with more

emotional support,
create greater opportunities for them to build their

competences and give them
more positive feedback than under induced low expectations.

Differential care that promotes in patients different
levels of self-efficacious
ness and skill in managing health-related behavior can

exert stronger impact on
the trajectories of health functioning than simply

conveying prognostic informa
tion. The effects of verbal prognostications alone may be

short-lived if they are
repeatedly disconfmned by personal experiences. But a sense

of personal effi
cacy rooted in enhanced competencies fosters functional

attainments that create
their own experiential validation. Clinical transactions

operate bidirectionally to
shape the course of change. The functional improvements

fostered by positive
expectancy influences further strengthen clinicians'

beneficial expectations and
their sense of efficacy to aid progress. In contrast,

negative expectations that
breed functional declines can set in motion a downward

course of mutual
discouragement. Medical conditions that produce severe

permanent impairments can be
devastatingly demoralizing to patients and their families.

Patients have to reorga
nize their perspective to learn alternative ways of

regaining as much control as
possible over their life activities. Goals need to be

restructured in functional
terms that capitalize on remaining capacities (Ozer, 1988).

Focus on achieve
ment of functional improvements rather than on degree of

organic impairments
helps to counteract self-demoralization. Making difficult

activities easier by
breaking them down into graduated subtasks of attainable

steps helps to prevent
self-discouragement of rehabilitative efforts and enhances

functional attainments. CONCLUDING REMARKS The converging
lines of evidence reported in this chapter indicate that
the
self-efficacy mechanism plays an influential role in

mediating the impact of psy
chosocial factors both on biological systems that

interrelatedly alter physical
functioning and on health habits that prevent or mitigate

pathogenic conditions.
The value of a psychological theory is judged not only by

its explanatory and
predictive power, but also by its operational power to

effect change in human
functioning. Social cognitive theory provides prescriptive

specificity on how to
enhance people's sense of personal efficacy in ways that

promote their health and
Ader, R., & Cohen, N. (1981). Conditioned

immunophannacologic responses. In R.
Ader (Ed.), Psychoneuroimmunology (pp. 281-319). New York:

Academic Press.
Ader, R., & Cohen, N. (1985). CNS-immune system

interactions: Conditioning phe
nomena. The Behavioral and Brain Sciences, 8, 379-394.
Ahles, T. A., Blanchard, E. B., & Leventhal, H. (1983).

Cognitive control of pain:
Attention to the sensory aspects of the cold pressor

stimulus. Cognitive Therapy and
Research, 7, 159-178.
Ajzen, 1., & Fishbein, M. (1980). Understanding attitudes

and predicting social
behavior. Englewood Cliffs, NJ: Prentice-Hall.
Ajzen, I., & Madden, T. 1. (1986). Prediction of

goal-directed behavior: Attitudes, in
tentions, and perceived behavioral control. Journal of

Experimental Social

Bandura, A. (1988a). Self-efficacy conception of anxiety.

Anxiety Research, 1,77-98.
Bandura, A. (1988b). Self-regulation of motivation and

action through goal systems.
In V. Hamilton, O. H. Bower, & N. H. Frijda (Eds.),

Cognitive perspectives on emo
tion and motivation (pp. 37-61). Dordrecht: Kluwer Academic

Publishers.

promoting behavior. In 1. Madden, IV, (Ed.), Neurobiology

of learning, emotion and
affect (pp. 229-270). New York: Raven.

Bandum, A. (1992). Psychological aspects of prognostic
judgments. In R. W. Evans,
D. S. Baskin, & F. M. Yatsu (Eds.), Prognosis in

neurological disease. New York:
Oxford University Press (in press).
Bandum, A., Cioffi, D., Taylor, C. B., & Brouillard, M. E.

(1988). Perceived self
efficacy in coping with cognitive stressors and opioid

activation. Journal of
Bandura, A., O'Leary, A., Taylor, C. B., Gauthier, J., &

Gossard, D. (1987). Perceived
self-efficacy and pain control: Opioid and nonopioid

mechanisms. Journal of

Microanalysis of action and fear arous
al as a function of differential levels of perceived

self-efficacy. Journal of Personality
Bandura, A., Taylor, C. B., Williams, S. L, Mefford, I.

N., & Barchas, J. D. (1985).

Consulting and Clinical Psychology, 53,406-414.
Baron, M., Dutil, E., Berkson, L., Lander, P., & Becker, R.
(1987). Hand function in
the elderly: Relation to osteoarthritis. Journal of

Rheumatology, 14, 815-819.

threat seriousness and person
al efficacy upon intentions and behavior. Journal of

Applied Social Psychology, II,
401-415.
Becker, M. H., & Maiman, L. A. (1983). Models of

health-related behavior. In D.
Mechanic (Ed.), Handbook of health, health care, and the

health professions (pp.
539-568). New York: The Free Press.
Bernier, M., & Avard, J. (1986). Self-efficacy, outcome and

attrition in a weight reduc
tion program. Cognitive Therapy and Research, 10, 319-338.
Blittner, M., Goldberg, J., & Merbaum, M. (1978). Cognitive

self-control factors in the
reduction of smoking behavior. Behavior Therapy. 9,

553-561.
Brod, M. I., & Hall, S. M. (1984). Joiners and non-joiners

in smoking treatment: A
comparison of psychosocial variables. Addictive Behaviors,

9, 217-221.
Cioffi, D. (1991). Beyond attentional strategies: Advancing

the study of somatic inter
pretation. Psychological Bulletin, 109,25-41.
Coe, C. L., & Levine, S. (1991). Psychoimmunology: An old

idea whose time has
come. In P. R. Barchas (Ed.), Social physiology of social

relations. Oxford: Oxford
University Press (in press).
Coletti, G., Supnick, 1. A., & Payne, T. 1. (1985). The

smoking self-efficacy question
naire (SSEQ): Preliminary scale development and validation.

Behavioral
Assessment, 7,249-260.
Council, 1. R., Ahern, D. K., Follick, M. 1., & Kline, C.
L. (1988). Expectancies and
functional impairment in chronic low back pain. Pain,

33,323-331.
Crabtree, M. K. (1986). Self-efficacy beliefs and social

support as predictors of diabetic
self-care. Unpublished doctoral dissertation, University

of California, San Francisco.
Craig, K. D. (1986). Social modeling influences: Pain in

context. In R. A. Sternbach
(Ed.), The psychology of pain (2nd ed., pp. 67-95). New

York: Raven Press.

infant temperament, and par
enting self-efficacy: A mediational model of postpartum

depression. Child
DeBusk, R. F., Kraemer, H. C., & Nash, E. (1983). Stepwise

risk stratification soon
after acute myocardial infarction. The American Journal of

Cardiology, 12, 1161
] 166.
Desharnais, R., Bouillon, 1., & Godin, G. (1986).

Self-efficacy and outcome expecta
tions as determinants of exercise adherence. Psychological

Reports, 59, 1155-1 159.

Self-efficacy: The third factor be
sides attitude and subjective norm as a predictor of

Education Research, 3,273-282.
DiClemente, C. C., Fairhurst, S. K., Velasquez, M. M.,

Prochaska, 1. 0., Velicer, W. F.,
& Rossi, J. S. (1991, in press). The process of smoking
cessation: An analysis of
precontemplation, contemplation and preparation stages of

change. Journal of Con
sulting and Clinical Psychology.
Dienstbier, R. A. (1989). Arousal and physiological

toughness: Implications for mental
and physical health. Psychological Review, 96, 84-100.
Dolce, J. 1. (1987). Self-efficacy and disability beliefs

in behavioral treatment of pain.
Behaviour Research and Therapy, 25,289-300.
Dzewaltowski, D. A. (1989). Towards a model of exercise

motivation. Journal of Sport
and Exercise Psychology, 11,251-269.
Dzewaltowski, D. A., Noble, J. M., & Shaw, J. M. (1990).

Physical activity participa
tion: Social cognitive theory versus the theories of

reasoned action and planned
behavior. Journal of Sport and Exercise Psychology,

12,388-405.
Engel, G. L. (1977). The need for a new medical model: A

challenge for biomedicine.
Science, 196, 129-136.
Ewart, C. K., Stewart, K. 1., Gillilan, R E., & Kelemen,

M. H. (1986). Self-efficacy
mediates strength gains during circuit weight training in

men with coronary artery di
sease. Medicine and Science in Sports and Exercise, 18,

531-540.
Fanselow, M. S. (1986). Conditioned fear-induced opiate

analgesia: A competing
motivational state theory of stress analgesia. In D. D.
Kelly (Ed.), Stress-induced
analgesia. Annals of the New York Academy of Sciences

(Vol. 467, pp. 40-54). New
York: New York Academy of Sciences.
Fries, J. E (1989). Aging well. Reading, MA:

Addison-Wesley.
Fries,1. E, & Crapo, L. M. (1981). Vitality and aging:

Implications of the rectangular
curve. San Francisco: Freeman.
Fry, L., Mason, A. A., & Pearson, R. S. B. (1964). Effect

of hypnosis on allergic skin
responses in asthma and hay fever. British Medical

Journal, 5391, 1145-1148.
Fuchs, V. R (1974). Who shall live? Health. economics, and

social choice. New
York: Basic Books.
Fuchs, V. R (1990). The health sector's share of the gross

national product. Science,
247, 534-538.
Glass, D. c., Reim, 8., & Singer, J. (1971). Behavioral

consequences of adaptation to
controllable and uncontrollable noise. Journal of

Experimental Social Psychology, 7,
244-257.
Glynn, S. M., & Ruderman, A. 1. (1986). The development

and validation of an eating
self-efficacy scale. Cognitive Therapy and Research, 10,

403-420.
Goldberger, L., & Breznitz, S. (Eds.).(1982). Handbook of

stress: Theoretical and
clinical aspects. New York: Free Press.

Gortner, S. R., & Jenkins, L. S. (1990). Self-efficacy and
activity level following cardi
ac surgery. Journal of Advanced Nursing, 15, 1132-1138.
Gossop, M., Green, L., Phillips, G., & Bradley, B. (1990).

Factors predicting outcome
among opiate addicts after treatment. British Journal of

Clinical Psychology, 29,209
216.
Gracely, R. H., Dubner, R, Wolskee, P. 1., & Deeter, W. R.

(1983). Placebo and
naloxone can alter post-surgical pain by separate

mechanisms. Nature, 306, 264-265.
Grau, J. W., Hyson, R. L., Maier, S. F., Madden, 1., IV, &
Barchas, 1. D. (1981). Long
term stress-induced analgesia and activation of the opiate

system. Science, 213, 1409
1411.
Holahan, C. K., & Holahan, C. 1. (1987a). Self-efficacy,

social support, and depression
in aging: A longitudinal analysis. Journal of Gerontology,

42,65-68.
Holahan, C. K., & Holahan, C. 1. (1987b). Life stress,

hassles, and self-efficacy in
aging: A replication and extension. Journal of Applied

Social Psychology, 17,574
592.
Holman, H., Mazonson, P., & Lorig, K. (1989). Health

education for self-management
has significant early and sustained benefits in chronic

arthritis. Transactions of the
Association of American Physicians, 102,204-208.

Holroyd, K. A., & Creer, T. L. (t 986). Self-management of
chronic disease. New York:
Academic Press.
Holroyd, K. A., Penzien, D. B., Hursey, K. G., Tobin, D.

L., Rogers, L., Holm, 1. E.,
Marcille, P. 1., Hall, J. R., & Chila, A. G. (1984).

Change mechanisms in EMG bio
feedback training: Cognitive changes underlying

improvements in tension headache.
Journal of Consulting and Clinical Psychology, 52,

1039-1053.
Jeffrey, R. W., Bjornson-Benson, W. M., Rosenthal, B. S.,

Lindquist, R. A., Kurth, C.
L., & Johnson, S. L. (1984). Correlates of weight loss and

its maintenance over two
years offollow-up among middle-aged men. Preventive

Medicine, 13, 155-168.
Jemmott, J. 8., Ill, & Locke, S. E. (1984). Psychosocial

factors, immunological media
tion, and human susceptibility to infectious diseases: How

much do we know? Psy
chological Bulletin, 95, 78-1 08.
Jones, R. A. (1977). Selfluljilling prophesies: Social,

psychological. and physiological
effects of experiences. Hillsdale, NJ: Erlbaum.
Jussim, L. (1986). Self-fulfilling prophecies: A

theoretical and integrative review.
Kahneman, E. (1973). Attention and effort. Englewood

Cliffs, NJ: Prentice-Hall.
Kanfer, R., & Zeiss, A. M. (t 983). Depression,

Kaplan, R. M., Atkins, C. J., & Reinsch, S. (1984).

Specific efficacy expectations

Psychology, 3. 223-242.
Kavanagh, D. J., & Wilson, P. H. (1989). Prediction of

outcome with a group version
of cognitive therapy for depression. Behaviour Research and

Therapy, 27, 333-347.
Kelley, D. D. (Ed.).(1986). Stress-induced analgesia.

Annals of the New York Academy
of Sciences (Vol. 467). New York: New York Academy of

Sciences.
Kiecolt-G1aser, J. K., & Glaser, R. (1987). Behavioral

influences on immune function:
Evidence for the interplay between stress and health. In

T. Field, P. M. McCabe, &
N. Schneiderman (Eds.), Stress and coping across

development (Vol. 2, pp. 189-206).
Krantz, D. S., Grunberg, N. E., & Baum, A. (1985). Health

psychology. Annual
Reviews in Psychology, 36,349-383.
Lazarus, R. S., & Folkman, S. (1984) .. New York: Springer.
Levine, 1. D., & Gordon, N. C. (t 984). Influence of the

method of drug administration
on analgesic response. Stress, appraisal, and coping.

Nature, 312, 755-756.
Levine,1. D., Gordon, N. c., & Fields, H. L. (1978,

September 23). The mechanism of
placebo analgesia. The Lancet, 654-657.

Levine, J. D., Gordon, N. C., Jones, R. T., & Fields, H.
L. (1978). The narcotic
antagonist naloxone enhances clinical pain. Nature,

272,826-827.
Levinson, R. A. (1986). Contraceptive self-efficacy: A

perspective on teenage girls'
contraceptive behavior. Journal of Sex Research, 22,

347-369. A. Bandura
Lin, M. D. (1988). Self-efficacy and perceived control:

Cognitive mediators of pain
tolerance. Journal of Personality and Social Psychology,

54, 149-160.
Locke, S., Ader, R., Besedovsky, H., Hall, N., Solomon,

G., & Strom, T. (Eds.). (1985).
Foundations ofpsychoneuroimmunology. Hawthorne, NY: Aldine.
Lorig, K, (1990, April). Self-efficacy: Its contributions

to the four year beneficialout
come of the arthritis self-management course. Paper

presented at the meeting of the
Society for Behavioral Medicine, Chicago, IL.
Lorig, K., Chastain, R. L., Ung, E., Shoor, S., & Holman,
H. (1989). Development and
evaluation of a scale to measure perceived self-efficacy

in people with arthritis.
Arthritis and Rheumatism, 32, 37-44.
Lorig, K, Seleznick, M., Lubeck, D., Ung, E., Chastain, R.

L., & Holman, H. R. (1989).
The beneficial outcomes of the arthritis self-management

course are not adequately
explained by behavior change. Arthritis and Rheumatism,

32, 91-95.
Love, S. Q., Ollendick, T. H., Johnson, C., & Schlezinger,

S. E. (1985). A preliminary
report of the prediction of bulimic behavior: A social
learning analysis. Bulletin of
the Society of Psychologists in Addictive Behavior, 4,

93-101.
Maccoby, N., & Solomon, D. S. (1981). Heart disease

prevention: Community studies.
In R. E. Rice & W. J. Paisley (Eds.), Public communication

campaigns (pp. 105-125).
Beverly Hills, CA: Sage Publications.
MacLennan, A. 1., Drugan, R. C., Hyson, R. L., Maier, S.

F., Madden, J., IV, &
Barchas, 1. D. (1982). Corticosterone: A critical factor in

an opioid form of stress
induced analgesia. Science, 215,1530-1532.
Maibach, E., Flora, J., & Nass, C. (1991). Changes in

self-efficacy and health behavior
in response to a minimal contact community health campaign.

Health Communica
tion, 3. 1-15.
Maier, S. F. (1986). Stressor controllability and

stress-induced analgesia. In D. D.
Kelly (Ed.), Stress-induced analgesia. Annals of the New

York Academy of Sciences
(Vol. 467, pp. 55-72). New York: New York Academy of

Sciences.
Maier, S. F., Laudenslager, M. L., & Ryan, S. M. (1985).

Stressor controllability, im
mune function, and endogenous opiates. In F. R. Brush & 1.

B. Overmier (Eds.),
Affect, conditioning, and cognition: Essays on the

determinants of behavior (pp. 183

Major, B., Cozzarelli, c., Sciacchitano, A. M., Cooper, M.
L., Testa, M., & Mueller, P.
M. (1990). Perceived social support, self-efficacy, and

adjustment to abortion.
Manning, M. M., & Wright, T. L. (1983). Self-efficacy

expectancies, outcome expect
ancies, and the persistence of pain control in childbirth.

McAuley, E. (1991, in press). Understanding exercise

behavior: A self-efficacy per
spective. In G. C. Roberts (Ed.), Understanding motivation

in exercise and sport.
Champaign, IL: Human Kinetics.
McCann, B. S., Follette, W. C., Driver, 1. L., Brief, D.

J., & Knopp, R. H. (1988,
August). Self-efficacy and adherence in the dietary

treatment of hyperlipidemia.
Paper presented at the 96th Annual Convention of the

American Psychological
Association, Atlanta, Georgia.
McCaul, K. D., Glasgow, R. E., & Schafer, L. C. (1987).

Diabetes regimen behaviors:
Predicting adherence. Medical Care, 25,868-881.
McCaul, K. D., O'Neill, K., & Glasgow, R. E. (1988).

Predicting the performance of
dental hygiene behaviors: An examination of the Fishbein

and Ajzen model and self
efficacy expectations. Journal of Applied Social

Mcintyre, K. 0., Lichtenstein, E., & Mermelstein, R. J.
(1983). Self-efficacy and
relapse in smoking cessation: A replication and extension.

McKusick, L., Coates, T. J., Morin, S. F., Pollack, L., &

Hoff, C. (1990). Longitudinal
predictors of reductions in high risk sexual behaviors

among gay men in San
Francisco: The AIDS behavioral research project. American

Journal of Public
Health, 80, 978-983.
Mefford, I. N., Ward, M. M., Miles, L., Taylor, B.,

Chesney, M. A., Keegan, D. L., &
Barchas, J. D. (1981). Determination of plasma

catecholamines and free 3,4-dihy
droxyphenylacetic acid in continuously collected human

plasma by high performance
liquid chromatography with electrochemical detection. Life

Sciences, 28, 447-483.
Meyerowitz, 8. E., & Chaiken, S. (1987). The effect of

message framing on breast self
examination attitudes, intentions, and behavior. Journal

O'Leary, A. (1990). Stress, emotion, and human immune

function. Psychological
Bulletin, J08, 363-382.
O'Leary, A., Goodhart, F., Jemmott, L. S., &

Boccher-Lattimore, D. (1991). Predictors
of safer sexual behavior on the college campus: A social

cognitive analysis. Journal
of American College Health (in press).
O'Leary, A., Shoor, S., Lorig, K., & Holman, H. R. (1988).

A cognitive-behavioral
treatment for rheumatoid arthritis. Health Psychology, 7,

527-544.
Ozer, M. N. (1988). The management of persons with spinal

cord injury. New York:
Demos.
Palmblad, J. (1981). Stress and immunologic competence:

Studies in man. In R. Ader
(Ed.) Psychoneuroimmunology (pp. 229-257). New York:

Academic Press.
Puska, P., Nissinen, A., Salonen, J. T., & Toumilehto, J.

(1983). Ten years of the North
Karelia project: Results with community-based prevention

of coronary heart disease.
Scandinavian Journal of Social Medicine, 11, 65-68.
Reese, L. (1983). Coping with pain: The role of perceived

self-efficacy. Unpublished
doctoral dissertation, Stanford University, Stanford, CA.
Rosenstock, I. M., Strecher, V. J., & Becker, M. H.

(1988). Socialleaming theory and
the health belief model. Health Education Quarterly, 15,

175-183.
Sallis, 1. F., Haskell, W. L., Fortmann, S. P., Vranizan,

M. S., Taylor, C. B., &
Solomon, D. S. (1986). Predictors of adoption and

maintenance of physical activity
in a community sample. Preventive Medicine, /5,331-341.
Sallis, 1. F., Pinski, R. 8., Grossman, R. M., Patterson,

T. L., & Nader, P. R. (1988).
The development of self-efficacy scales for health-related

diet and exercise behav
iors. Health Education Research, 3, 283-292.
Schifter, D. E., & Ajzen, I. (1985). Intention, perceived

control, and weight loss: An
application of the theory of planned behavior. Journal of

Schneider, J. A., O'Leary, A., & Agras, W. S. (1987). The

role of perceived self
efficacy in recovery from bulimia: A preliminary

examination. Behaviour Research
and Therapy, 25,429-432.
Schwarzer, R. (1992). Self-efficacy in the adoption and

maintenance of health behav
iors: Theoretical approaches and a new model. In

Schwarzer, R. (Ed.), Self-efficacy:
Thought control of action. Washington, DC: Hemisphere

(this volume).
Shavit, Y., & Martin, F. C. (1987). Opiates, stress, and

immunity: Animal studies.
Shoor, S. M., & Holman, H. R. (1984). Development of an

instrument to explore psy
chological mediators of outcome in chronic arthritis.

Transactions of the Association
of American Physicians, 97, 325-331.
Sitharthan, T., & Kavanagh, D. 1. (1991). Role of

self-efficacy in predicting outcomes
from a programme for controlled drinking. Unpublished

manuscript, Drug and Alco
hol Department, Royal Prince Alfred Hospital, Camperdown,

Australia.
Slater, M. D. (1989). Social influences and cognitive
control as predictors of self
efficacy and eating behavior. Cognitive Therapy and

Research, /3,231-245.
Smith, G. R., & McDaniel, S. M. (1983). Psychologically

mediated effect on the de
layed hypersensitivity reaction to tuberculin in humans.

Psychosomatic Medicine, 45,
65-70.
Solomon, K. E., & Annis, H. M. (1990). Outcome and

efficacy expectancy in the pre
diction of post-treatment drinking behaviour. British

Journal of Addiction, 85, 659
666.
Stotland, S., Zuroff, D. c., Roy, M. (1991). Situational

dieting self-efficacy and short
term regulation of eating. Unpublished manuscript, McGill

University, Montreal,
Quebec, Canada.


husbands' cardiac capabilities soon
after clinically uncomplicated acute myocardial infarction.

American Journal of
Cardiology, 55,635-638.
Turk, D., Meichenbaum, D., & Genest, M. (1983). Cognitive

therapy of pain. New
York: Guilford.
Watkins, L. R., & Mayer, D. 1. (1982). Organization of

endogenous opiate and
nonopiate pain control systems. Science, 2/6, 1185-1192.

Levine, S., & Raska, K.

stressors on components of
the immune system. Journal of Personality and Social

Psychology, 59, 10821094.

Generalization of therapeutic chang
es in agoraphobia: The role of perceived self-efficacy.

Wilson, G. T., Rossiter, E., Kleifield, E. I., & Lindholm,

L. (1986). Cognitive-behav
ioral treatment of bulimia nervosa: A controlled

evaluation. Behavior Research and
Therapy, 24,277-288.
Winett, R. A., King, A. C., & Altman, D. G. (1989). Health

psychology and public
health: An integrative approach. Elmsford, NY: Pergamon.

10.4324 9781315800820 Previewpdf

Uploaded by

Copyright:

Available Formats

You might also like

10.4324 9781315800820 Previewpdf

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

10.4324 9781315800820 Previewpdf

Uploaded by

Copyright:

Available Formats

Self-Efficacy

This page intentionally left blank

LONDON AND NEW YORK

This edition published 2014 by Routledge

Routledge is an imprint o/the Taylor & Francis Group, an informa business

SELF-EFFICACY: Thought Control of Action

Copyright © 1992 by Hemisphere Publishing Corporation. All rights reserved.

Cover design by Michelle Fleitz.

Library of Congress Cataloging-in-Publication Data

I Self-Efficacy and Human Functioning

Exercise of Personal Agency Through the Self-Efficacy Mechanism

II Self-Efficacy and Human Development

Perceived Control: Motivation, Coping, and Development

III Self-Efficacy, Stress, and Emotions

Perceived Self-Efficacy and Phobic Disability

IV Self-Efficacy and Health Behaviors

Self-Efficacy in the Adoption and Maintenance of Health

V Self-Efficacy, Physical Symptoms, and Rehabilitation of

The Role of Physical Self-Efficacy in Recovery From

Author Index 395

Aart N. Mudde, Rijksuniversiteit Limburg, Department of Health Education,

The present volume, organized in five sections, covers a broad range of

performance is related to both emotions and self-related cognitions. All three

distinguished as decision-making strategies, interpersonal strategies, and self-

chronic disease and in the rehabilitation process. Self-efficacy improves changes

Los Angeles, September 1991 Ralf Schwarzer

The present chapter analyzes the influential role of perceived self-

The recent years have witnessed a resurgence of interest in self-referent phenom-

SELF ·EFFICACY CAUSALITY

LEVEL OF PERCEIVED SELF-EFFICACY

Figure 2 Mean levels of mathematical solutions achieved by students as a function of

LOW NO HIGH LOW NO HIGH

Figure 3 Mean changes induced in perceived self-efficacy by anchoring influences and

Figure 4 Mean level of physical stamina mobilized in competitive situations as a func-

LOW HIGH LOW HIGH

Figure 5 Mean changes in perceived self-efficacy induced by arbitrary normative com-

sociocognitive functioning is further documented in comparative tests of the

EFFICACY -ACTIV ATED PROCESSES

Self-beliefs of efficacy affect thought patterns that can enhance or undermine

A major function of thought is to enable people to predict the occurrence of

EFFICIENT USE OF ANALYTIC STRATEGIES

STRENGTH OF PERCEIVED SELF-EFFICACY

STRENGTH OF PERCEIVED SELF-EFFICACY

perceived self-efficacy, retards physical skill development and diminishes

.57 (.75) .37 (.55)

EFFICIENT USE OF ANALYTIC STRATEGIES

STRENGTH OF PERCEIVED SELF-EFFICACY

PHASES OF THE EXPERIMENT

Seeing oneself surpassed by similar social referents undermined perceived

Self-beliefs of efficacy playa central role in the self-regulation of motivation.

Figure 10 Schematic representation of conceptions of cognitive motivation based on

The self-efficacy mechanism of personal agency operates in all of these vari-

efficacy when contingencies are discriminatively structured so that no level of

solely by increased level of effort, self-discontent with substandard attainments is

to INCREASE IN PERFORMANCE MOTIVATION

Figure 11 Mean percent change in motivational level as a function of the number of

Many theories of motivation and self-regulation are founded on a negative

It is widely believed that misjudgment breeds dysfunction. The functional